introduction to radiobiology lesson 6 - …milotti/didattica/introductory... · edoardo milotti -...

Introduction to RadiobiologyLesson 6

Master of Advanced Studies in Medical Physics A.Y. 2017-18

Edoardo MilottiPhysics Dept. – University of Trieste

Edoardo Milotti - Radiobiology 2

Radiobiological knowledge is used to optimize treatment

In this lesson we consider some concepts associated with radiobiology and related to treatment optimization

1. Fractionation

2. The 4 R’s (5R’s) of radiobiology

3. Dose-volume histograms (DVH) and isodose curves

4. Equivalent Uniform Dose (EUD)

5. Optimization (basic concepts of treatment plans)


1. Fractionation

Fractionation of radiation treatment specifies how to split dose over a period of weeks rather than in a single session so that the treatment results in a better therapeutic ratio.

To achieve the desired level of biological damage the total dose in a fractionated treatment is much larger than that in a single treatment.


A simple Poisson model of the surviving fraction has an exponent that is proportional to the dose. Fractionation approaches this linear behavior.

S(D) = e�D/D0

= 10D log10 e/D0

= 10�D/D10

lnS(D) = �D/D0

D0 =�D

� ln(1/S(D))

D10 = D0/ log10 e ⇡ 2.3D0

D10 =�D

� log10(1/S(D))<latexit sha1_base64="yzBsh7RF/1JiSxPNaWI0FKBthjM=">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</latexit><latexit sha1_base64="yzBsh7RF/1JiSxPNaWI0FKBthjM=">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</latexit><latexit sha1_base64="yzBsh7RF/1JiSxPNaWI0FKBthjM=">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</latexit><latexit sha1_base64="yzBsh7RF/1JiSxPNaWI0FKBthjM=">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</latexit>


The algebra of fractionation, using the linear-quadratic law

Survival probability with n doses D

The corresponding biological effect is

[S(D)]n

E = � ln[S(D)]n = �n lnS(D)

= n(↵D + �D2)

= ↵(nD)

✓1 +

D

↵/�

◆

total doserelative effectiveness


BED =E

↵= (nD)

✓1 +

D

↵/�

◆

biologically effective dose total dose

relative effectiveness

The relative effectiveness is always > 1, therefore the biologically effective dose is always greater than the total dose.


Response to fractionation varies with tissue, fractionation spares late responding tissues

16128400.01

.1

1

Dose (Gy)

S.F.

When a/b is high (>6Gy) the survival curve is almost exponential, when a/bis low (1-4Gy) the shoulder is wide

2016128400.01

.1

1

Dose (Gy)

S.F.Early-responding tissues

Late-responding tissues

Single dose early-resp. tissues

Single dose late-resp. tissues

Fractionated dose early-resp. tissues

Fractionated dose late-resp. tissues


An equation for BED that includes tumor repopulation

After a ”kickoff time” Tk, tumor cells start proliferating again, therefore the tumor population after treatment has changed by the total factor

where Tp is the tumor cells’ duplication time. Taking logarithms, we find

N(T )/N0 = [S(D)]n2(T�Tk)/Tp

n ln[S(D)] +T � Tk

Tp/ ln 2= �↵nD

✓1 +

D

↵/�

◆+

T � Tk

Tp/ ln 2


n ln[S(D)] +T � Tk

Tp/ ln 2= �↵nD

✓1 +

D

↵/�

◆+

T � Tk

Tp/ ln 2

BED(D,n, T ) = (nD)

✓1 +

D

↵/�

◆� T � Tk

↵Tp/ ln 2

= BED(D,n)� T � Tk

↵Tp/ ln 2


Example, conventional treatment:

30F x 2Gy/6 weeks

39% difference between early- and late-responding

BED(early) = (nD)

✓1 +

D

↵/�

◆

= (60 Gy)

✓1 +

2

10

◆

= 72 Gy10

BED(late) = (60 Gy)

✓1 +

2

3

◆

= 100 Gy3


Example, hyperfractionation:

70F x 1.15 Gy twice daily/7 weeks


BED(early) = (nD)

✓1 +

D

↵/�

◆

= (80.5 Gy)

✓1 +

1.15

10

◆

= 89.8 Gy10

BED(late) = (80.5 Gy)

✓1 +

1.15

3

◆

= 111.4 Gy3


The basic aim of hyperfractionation is to further separate early and late effects.

The overall treatment time remains conventional at 6 to 8 weeks, but because two fractions per day are used, the total number of fractions is 60 to 80.

The dose must be increased because the dose per fraction is decreased.

Early reactions may be increased slightly, tumor control improved, and late effects greatly reduced.

Hyperfractionation has been shown in randomized clinical trials of head and neck cancer to improve local tumor control and survival with no increase in acute or late effects.


CHART stands for continuous hyperfractionated accelerated radiation therapy.

The protocol consists of 36 fractions over 12 days (three fractions per day) to a total dose of 50.4 to 54 Gy.

Tumor control is maintained because of the extreme acceleration of treatment time; late effects are not increased and may even decrease because of the low dose; the acute effects are severe, but their peak occurs after completion of treatment, so patient compliance is not prejudiced.


Example, CHART:

36F x 1.5 Gy (3F/day)/12 days


BED(early) = (nD)

✓1 +

D

↵/�

◆

= (54 Gy)

✓1 +

1.5

10

◆

= 62.1 Gy10

BED(late) = (54 Gy)

✓1 +

1.5

3

◆

= 81 Gy3


Isoeffect equationwhen two fractionation strategies have the same BED we find

For comparison purposes it is useful to define the Equivalent Dose at 2Gy:

equivalent dose with 2Gy fractions

total dose delivered in d Gy fractions

EQD2Gy = Dd+ ↵/�

2Gy + ↵/�

D1 [1 + d1/(↵/�))] = D2 [1 + d2/(↵/�)]


Summary of conventional wisdom

• The LQ model satisfactorily describes the relationship between total isoeffectivedose and dose per fraction over the range of dose per fraction from 1 Gy up to 5–6 Gy.

• The α /β ratio describes the shape of the fractionation response: a low α /β (0.5–6 Gy) is usually characteristic of late-responding normal tissues and indicates a rapid increase of total dose, with decreasing dose per fraction and a survival curve for the putative target cells that is significantly curved.

• A higher α /β ratio (7–20 Gy) is usually characteristic of early-responding normal tissues and rapidly-proliferating carcinomas; it indicates a less significant increase in total dose with decreasing dose per fraction and a less curved cell-survival response for the putative target cells.


Summary of conventional wisdom (ctd.)

• The EQD2 formulae provide a simple and convenient way of calculating isoeffectiveradiotherapy schedules, based on the LQ model. Tolerance calculations always require an estimate of the α /β ratio to be included.

• For short interfraction intervals, a correction may be necessary for incomplete repair. When using the EQD2 formulae to calculate schedules with multiple fractions per day or continuous low dose rate, an estimate of the repair halftime must also be included.

• The basic LQ model is appropriate for calculating the change in total dose for an altered dose per fraction, assuming the new and old treatments are given in the same overall time. For late reactions it is usually unnecessary to modify total dose in response to a change in overall time, but for early reactions (and for tumourresponse) a correction for overall treatment time should be included. Although the effect of time on biological effect is complex, simple linear corrections have been shown to be of some value.


QUANTEC guidelinesThe Quantitative Analysis of Normal Tissue Effects in the Clinic (QUANTEC) guidelines are a recent effort to review and summarize normal tissue toxicity, which may suggest dose-volume treatment planning guidelines and likely reduce the rates of side effects.

The primary goal is to provide a simple set of data to be used by the busy community of practitioners of radiation oncology, physicists and dosimetrists.

The second goal is to provide reliable predictive models of the relationships between dose-volume parameters and normal tissue complications to be used in the planning of radiation therapy.

The results of this large study can be found on this webpage

http://aapm.org/pubs/QUANTEC.asp

Note that these guidelines are not final and shall certainly be revised in the future as new data become available.

http://aapm.org/pubs/QUANTEC.asp


2. The 4 R’s (5 !!!) of radiotherapy: a radiobiological rationale for fractionated radiotherapy

Radiobiological mechanisms that impact on the efficacy of radiotherapy.A summary list of what is important in radiotherapy (introduced by Withers in 1975)

A. Repair

B. Redistribution of cells within the cell cycle

C. Repopulation

D. Reoxygenation

… and

E. Radiosensitivity (the new, 5th R)


A. RepairThe repair of sublethal damage must be taken into account

• because it affects the tolerance of healthy tissue to radiotherapy (allowing cells to repair we can continue a treatment that should otherwise be interrupted)

• because tumor cells often have a reduced ability to repair damage, e.g., when they have a mutated P53 gene

When considering repair one must keep into account the mean repair time of healthy tissue – e.g., the spinal cord tissue has a slow mean repair time of about 4 hours, and this means that daily doses must have at least this separation to spare that tissue.

Dose rate must also be taken into account: too low a dose rate means that both healthy and tumor tissues can start repair during a session.


B. RedistributionProliferating cells have different radiosensitivities. After a session more of the cells in the S phase survive, and waiting for a redistribution of cells in different phases helps in killing them.

A low dose rate means that redistribution can take place during a session, and this should be taken into account.


C. RepopulationRepopulation takes place both in healthy and in diseased tissues.

Usually healthy early-responding tissues begin repopulation at about 4 weeks into treatment. Prolonging treatment over 4 weeks means a reduced early radiotoxicity for these tissues. This is not relevant for late-responding tissues.

At least some tumors display accelerated repopulation after 4-5 weeks into treatment. This means that this repopulation must be countered in long treatments.


RAT TUMOUR RESPONSE

A comparison of the growth rates ofthese two cell lines both in the animaland in culture is of interest. Although theR1/LBL line grows faster in vivo than theR2D2 line (TD ranging from 4-8-7 daysover a 20 day period cf 5-8-10 days) thereverse is true in vitro where the culturedR2D2 cells have a shorter doubling time,14 h cf 18 h for R1/LBL. Thus, there isno correlation between the two cell linesin their growth rates in vivo and in vitro.

Studies have recently been reported onthe growth delay observed for tumoursof the R1/LBL line after helium-ion andneon-ion irradiation (Curtis et al., 1978).Fig. 1 shows the volume response after

various doses of 220-kV X-rays as anexample of the type of data obtained.The radiation-induced growth delay isdetermined from the tumour volumeresponse data by calculating the differ-ence between the times for the irradiatedand the control tumours to double involume.

In Fig. 2 tumour growth delays areplotted as a function of dose for carbon-,neon- and argon-ion beams, and are com-pared with the growth delay curveobtained following exposure to X-rays.RBE's for different growth delays e.g.50 days (RBE5o values) and their standarddeviations are easily calculated from these

TABLE I.-RBE's for growth delay and tumour cureRadiation modality Initial energy Tumour position RBE20

12C 400 MeV/u 4 cm extended peak 2 8+0 720Ne 400, 425 MeV/u 4 cm extended peak 2 9 + 0 * 740Ar 570 MeV/u 4 cm extended peak 3 0 + 0* 612C 400 MeV/u plateau 1*3 + 0*320Ne 400, 425 MeV/u plateau 1 *7 + 04

15 MeV neutrons* 3. 3* From Barendsen & Broerse, 1969.t Extrapolated value.t RBE for TCD 90/120, calculated from extrapolation of cell survival data.

0)E

0

-

n

00)

.NtoE0z

RBE502-3+0-42-6+0-52-5+0-31*3+0-21*8+0-3

3. it

RBETCD50/1802-3+0.33-1+0-6

2-8:

Days post-irradiationFIG. 1.-Volumes of R1/LBL tumours are plotted as a function of time for controls and for tumours

receiving graded doses of 220-kV X-rays. The volumes have been normalized to unity on the dayof irradiation. Numbers in parentheses represent the number of tumours exposed to each radiationdose. Error bars represent one standard error of the mean.

267

from

Cur

tis &

Tenf

orde

: “As

sess

men

t of t

umou

rres

pons

e in

a ra

t rha

bdom

yosa

rcom

a”, B

r. J.

Canc

er 4

1 (1

980)

, 266


D. ReoxygenationMany tumor tissues are hypoxic, and this protects tumor cells from radiation because of the Oxygen Effect. Therefore one useful strategy consists in helping oxygen diffuse through tissues.

Reoxygenation can be achieved by killing cells closer to blood vessels, so that oxygen penetrates more deeply into the tumor tissue, and also using growth factors that reestablish a healthier, more regular vascularization in the tumor tissue (e.g, VEGF).


RAT TUMOUR RESPONSE

R2D2: 225kV X-roys R2D2: Neon Spread-out Brogg Peak

Dose (Groys) Dose (Groys) Dose (Grays)

FIG. 4.-Oxic survival curves are compared for R2D2 cells irradiated in vitro as a suspension andin vivo within a solid tumour. The survival curve for oxygenated cells in vivo was calculated fromthe measured survival curves for tumours irradiated in situ in air-breathing rats (data pointsshown) and in nitrogen-asphyxiated rats.

can be calculated. These calculated curvesare shown as dashed lines in Fig. 4, fortumours irradiated in situ in air-breathinganimals. In each figure the oxic in vitrosurvival curve for this cell line irradiatedin suspension is shown for comparison.For the carbon-ion and neon-ion beams,the suspension cultures were placed at thesame position as the tumours in theextended peak region of each beam. It isclear that the survival curves calculatedfor oxygenated cells in situ imply lessradiosensitivity (or more repair?) than thesame cells irradiated in suspension. Thisfinding is in accord with the observationsof Durand & Sutherland (1972) on thelower survival of single cells as comparedwith cells growing in multicellular spher-oids.A small series of experiments was per-

formed to obtain data on tumour growthdelay for the R2D2 cell line, thus allowing

TABLE II. RBE's for cell killing andgrowth delay

Radiation modalityPeak carbon ionsPeak neon ions15 MeV neutrons*

RBEo.11 9+0-33-1 +0-6

2-7

RBE50 day s

2-3+0-22-9+0-232- t

* From Barendsen & Broerse, 1969.t Extrapolated value.

a direct comparison of RBE's for growthdelay and cell survival with the same cellline. This comparison is shown in Table IIalong with the results obtained earlier byBarendsen & Broerse (1969) for 15 MeVneutrons. The RBE's for 10% survivalafter culture in vitro and for regrowthdelay in situ are closely similar.The general picture emerging from these

data is that, for both of the heavy charged-particle beams tested, RBE's for thevarious end points (in situ and in vitro cellsurvival, TCD50 and growth delay) arequite comparable. On the other hand,there are significant differences betweenin situ and in vitro survival curves and thedifferent cell lines used in these studieshave widely differing clonogenicity andgrowth properties. It is concluded thatthere is currently not enough known of thepost-irradiation cellular dynamics to pre-dict the ultimate fate of tumours left inthe animals following irradiation. Upuntil now, attempts to develop an all-inclusive model for the radiation responseof the rhabdomyosarcoma tumour systembased on cell-kinetic parameters andusing in vitro survival curves have metwith only limited success in predictingvolume response (Curtis et al., 1973). Moremust be learned about both the short-term

269

c

CY

I

from

Cur

tis &

Tenf

orde

: “As

sess

men

t of t

umou

rres

pons

e in

a ra

t rha

bdom

yosa

rcom

a”, B

r. J.

Canc

er 4

1 (1

980)

, 266


E. RadiosensitivityRadiosensitivity differs in different cell types, and this factor must be included in the therapeutic strategy.

Radiosensitivity can be enhanced in tumor cells with proper sensitizing chemicals.


3. Isodose curves and dose-volume histograms (DVH)

Dose is distributed in space and both tumor tissue and normal tissue are affected.

For this reason it is important to characterize the dose received by both tumor tissue and normal tissue in a quantitative way.


sione, indicazione al trattamen-to, follow up), sia le procedurecorrelate agli aspetti tecnici dellaterapia devono essere soggetti adattenti controlli di qualità;

• si raccomanda che ogni Centropartecipi regolarmente a inter-confronti dosimetrici esterni. Èstato infatti dimostrato che gliinterconfronti dosimetrici, o piùampie verifiche esterne della do-simetria e del funzionamentodelle apparecchiature, sono mol-to efficaci nell’evidenziare dovepossano potenzialmente insor-gere problemi.L’attività del Laboratorio di Fi-

sica in questo campo ha avuto ini-zio nel 1995 per far fronte al primodei due punti indicati, con la co-stituzione di Gruppi di studio sultema della GQR, a cui partecipanostudiosi indicati dalle associazionidei radioterapisti oncologi, dei fisi-ci specialisti e dei tecnici di ra-dioterapia, cioè proprio quel-le figure profes-sionali princi-palmente coin-volte nel tratta-mento radiote-rapico. Su que-sto fronte è atti-vo in primo luo-go l'ESTRO(European So-ciety for Thera-peutic Radiology and On-cology) con il suo "QualityAssurance Committee" che prov-vede a emanare linee guida genera-li. L’attività dei Gruppi ISS è rivol-ta a identificare gli aspetti sui qua-li sviluppare le suddette raccoman-dazioni e regole applicabili a livel-lo nazionale, in base a quanto in-dicato dal documento dell’ESTRO(8) elaborato nell'ambito del pro-gramma "Europe Against Cancer",e in analogia a quanto fatto daglialtri Paesi europei. La GQR, infat-ti, non è più concepita solo comecontrollo e taratura degli strumen-ti adoperati, ma secondo criteri diGQ globale che coprano tutti gli

aspetti delle procedure collegate al-la diagnosi, al trattamento e alfollow up. È necessario, per-

tanto, procederecon un’attivitàmultidisciplinarein modo da co-prire sia gli aspet-ti tecnici della te-rapia - come leprocedure di irra-diazione - che iparametri correla-ti al paziente,

quali diagnosi, decisioni rela-tive al trattamento, ecc.Tra le ricadute positive dell’in-

troduzione di un sistema di qua-lità vi è quello di garantireuna miglioreomogeneità nelmodo in cui unostesso protocolloterapeutico vie-ne applicato inCentri diversi;viene così mi-gliorata la pro-babilità di dimo-strare l’esistenza di diffe-renze tra le modalità tera-peutiche confrontate nello studio,

riducendo le variabilità legate a di-versità tra i Centri partecipanti.L’esistenza di procedure di GQ inuno studio multicentrico è quindioggi considerata requisito essen-ziale per la validazione dei risulta-ti riportati.

Per una buona pratica clinica,l’attuale legislazione (9) sottolineache, in tutte le articolazioni orga-nizzativo-funzionali, sia favoritol’utilizzo di linee guida predispostedalle società scientifiche o da grup-pi di esperti nelle varie branche spe-cialistiche.

I suddetti Gruppi di studio ISShanno avviato da diversi anni ini-

ziative relative alla GQR, or-ganizzando corsi-dibattito ed

elaborando lineeguida su questotema. Il coinvol-gimento dell’ISSin questo settoreè stato ribaditoanche nel nuovoregolamento diorganizzazionedell’ente (10) do-

ve è riportato, tra i compiti isti-tuzionali, quello di svolgere attivi-tà di consulenza per la tutela dellaN

ot

Ist

Sup

erSa

nit

à 20

02; 1

5(4

)

L’ISS ha elaborato linee guide generali

sulla garanzia di qualitàin radioterapia

Un sistema di qualità in radioterapia

garantisce omogeneitànell’applicazionedi un protocollo

terapeutico

Figura 5 - Esempio di curve di isodose bidimensionali calcolate per le irra-diazioni di un tumore (liposarcoma retroperitoneale) situato in prossimità diorgani critici (reni e midollo spinale). La regione tratteggiata rappresenta ilvolume bersaglio pianificato (PTV), mentre con la sigla OAR sono indicati gliorgani critici

Example of two-dimensional isodose curves in the treatment of retroperitoneal liposarcoma, close to critical organs – kidneys and spinal cord.

PTV = Planned Target VolumeOAR = Organ At Risk


logo

Treatment Planning Topics Photon Therapy 3D-conformal Radiotherapy

Radiation Dose

Dose is the measure of energydeposited in medium by ionizingradiation per unit mass

gray (Gy) = JouleKilogram

Dose deposited in patient ismeasured using a fine cubical grid

cubes are called voxels

8 / 39

Dose deposited in patient is measured using a fine cubical grid, and the cubes are called voxels.


logo


Dose Distribution Visualization

Visualizing the dose distributiondose-volume histogram (DVH)isodose linesdose-wash diagram

9 / 39

logo


Dose Distribution Visualization

Visualizing the dose distributiondose-volume histogram (DVH)isodose linesdose-wash diagram

9 / 39

Dose-volume histograms are cumulative distributions of the voxels receiving at least the given dose.

Differential dose-volume histograms are also used (fraction of the voxels receiving exactly the given dose).


4. Equivalent Uniform Dose (EUD)

According to Niemierko (who introduced the concept in 1997),

“For any dose distribution, the corresponding Equivalent Uniform Dose EUD is the dose in Gy, which, when distributed uniformly across the target volume, causes the survival of the same number of clonogens.”

In the discussion of EUD, “it is assumed that an irradiated tumor is composed of a large number of independent clonogens, and that random killing of the clonogens is well described by Poisson statistics. The binary response – control or failure – of an irradiated tumor is assumed to be determined by the expected number of surviving clonogens. Therefore, two different target dose distributions are equivalent if the corresponding expected number of surviving clonogens are equal.”

from Niemierko, Med. Phys. 24 (1997) 103


Example with a Poisson cell kill model

Surviving fraction for a generic dose D and for a reference dose Dref

The surviving fraction can be rewritten as an explicit function of the reference dose

S(D) = exp(�D/D0)

S(Dref) = exp(�Dref/D0)

lnS(Dref) = �Dref/D0 ! D0 = � Dref

lnS(Dref)

lnS(D) = �D/D0 = DlnS(Dref)

Dref! S(D) = (S(Dref))

D/Dref


Now assume that there are N cells uniformly scattered in a volume V, which is subdivided in subvolumes Vi which receive each a dose Di. Then the number of cells that survive in the whole volume is the sum of the cells that survive in subvolumes Vi

Therefore the total surviving fraction is

X

i

ni =N

V

X

i

Vi S(Di) =N

V

X

i

Vi (S(Dref))Di/Dref

1

N

X

i

ni =X

i

Vi

V(S(Dref))

Di/Dref =X

i

vi (S(Dref))Di/Dref

We would obtain the same surviving fraction

with an equivalent uniform dose EUD such that


S̄ =X

i

vi (S(Dref))Di/Dref

S̄ = (S(Dref))EUD/Dref

! EUD = Drefln S̄

lnS(Dref)= Dref

lnP

i vi S(Dref)Di/Dref

lnS(Dref)


This holds for the simple Poisson-model surviving fraction. More complex cases are treated in the paper by Niemierko:

• absolute volume effect

• nonuniform spatial distribution of clonogens

• dose-per-fraction effect (using the LQ model)

• proliferation effect

• inhomogenity of patient population


5. Optimization (basic concepts of treatment plans)

We optimize a treatment by

• maximizing damage to tumor tissue• minimizing damage to normal tissue

This is a complex process that requires numerical solutions.

In the following slides we analyze a simple example that utilizes Monte Carlo simulation to analyze the effects of an IMRT (Intensity-Modulated Radiation Therapy) treatment (IMRT is an improved version of the 3D-treatment).


In this example the radiation is delivered by beams with the same Gaussian intensity modulation (this kind of intensity modulation is not realistic, it is just part of this specific example)

beam profile

position (a.u.)-2 -1 0 1 2

0.0

0.2

0.4

0.6

0.8

1.0


Simulation target: glioblastoma multiforme

Glioblastoma multiforme (GBM) is the most common and malignant brain tumor found in human beings, accounting for approximately 52% of all functional tissue brain tumor cases and 20% of all intracranial tumors.

GBM is comprised of heterogeneous groups of neoplasms that proliferate through various parts of the central nervous system. Although it is the most prevalent form of primary brain tumor, only 2-3 cases per 100,000 people in the Europe and North America are reported annually. However, the prognosis for patients afflicted with GBM is extremely poor, and is eventually fatal in the vast majority of cases.

(from https://sites.google.com/site/whatisglioblastomamultiforme/pathophysiology)

https://sites.google.com/site/whatisglioblastomamultiforme/pathophysiology


0 100 200 300 4000.0

0.2

0.4

0.6

0.8

1.0

Dose (Gy)

TCPandNTCP

FIG. 2. TCP curves (solid black lines) for various GBM histological types12 and NTCP curve

(dashed red line) for brain tissue vs. dose D (Gy). The TCP curves have been drawn taking the

linear extrapolation14 of the LQ model with the ↵ and � parameters listed in Ref. 12. The NTCP

curve has been drawn with partial volume v = 5%.

maximizing the TCP.

To compute the TCP I approximate human cells with spheres with the average radius

R = 10 µm – see, e.g., Ref. 15, which is an excellent source of important biological numbers

like cellular sizes. This means that the average cell volume is about 4000 µm3 = 4⇥10�15 m3,

and – with a relative mass density very close to 1 – the average cell mass is about 4⇥10�12 kg.

The corresponding cell density is about 2.5⇥ 108 cells/cm3. Then, a solid tumor with about

109 cells20 has total volume ⇡ 4⇥ 103 mm3 and mass ⇡ 4⇥ 10�3 kg, which I attribute to a

disc 1 cm high and with a radius of about 1.1 cm. I assume that the radius of the head is

9 cm.

I start the trial-and-error by delivering a total dose D = 100 Gy, which is a common

value in many treatments, and corresponds to a total energy of 0.4 J released in the disc-

shaped tumor mass. The photon energy in many treatments ranges from about 1 MeV to

somewhat more than 10 MeV (see, e.g., Ref. 16), so that the number of photons absorbed

in the tumor volume ranges from about 2.5⇥ 1011 to about 2.5⇥ 1012. This is a very large

number of photons, and it is easy to anticipate that it is not possible to simulate all of them,

one has to simulate a lower dose, and then scale the distribution of the absorbed photons

to the higher dose. While this is perfectly acceptable if one is content with estimates of

average values, it also means that fluctuations observed in the simulation runs are mostly

artifacts due to the reduced number of photons. If the number of photons of energy E�

5

TCP curves (solid black lines) for various GBM histological types and NTCP curve (dashed red line) for brain tissue vs. dose D (Gy), for 109 cells (volume about 4 cm3). The TCP curves have been drawn taking the linear extrapolation of the LQ model with the α and β parameters listed in the literature. The NTCP curve has been drawn with partial volume v = 5%.


Example distribution with 3 beams

Each dot represents the position of one absorbed photon. The local dot density is proportional to the local dose. The photon beams undergo exponential attenuation, and there is a corresponding energy absorption in tissue.


Isodose curves

From voxels to isodose curves


0 20 40 60 80 100 1200.0

0.2

0.4

0.6

0.8

1.0

Dose (Gy)

Relativefrequency

0 20 40 60 80 100 1200.0

0.2

0.4

0.6

0.8

1.0

Dose (Gy)

FIG. 9. Dose-volume histograms (DVH) for the whole volume of the simulated head (left panel)

and for the planning target volume (right panel). DVH’s are empirical cumulative distributions of

dose that are often used in radiotherapy, but they are read o↵ di↵erently from usual cumulative

distributions. For instance, from the histogram on the right we find that about 70% of all voxels

receives a dose larger than 60 Gy, and that about 30% of all voxels receives a dose larger than 100

Gy. The histograms show here refer to the same MC simulation shown in figure 6.

The dose distribution in individual voxels and the isodose map are shown in figure 10. It

turns out that the change is not beneficial, the new configuration has roughly the same TPC

value and a larger NTCP value: TCP ⇡ 1.7% and NTCP ⇡ 16.7%.

In a final attempt I go back to the 3-beam configuration and try with wider Gaussian

beams to obtain a better coverage of the planning target volume, i.e., I use the parameters

• number of beams: 3

• beam angles: -60�, 0�, and 60�

• beam profile: Gaussian with common � = 2 cm

• total dose in the planning target volume: 200 Gy

This choice of parameters leads to a di↵erent distribution of dose, as shown in figure 11.

Again, the change does not significantly modify the TCP and NTCP: TCP ⇡ 1.9% and

NTCP ⇡ 86.1%. There is a minor gain in TCP and a very large increase in NTCP. The

situation is somewhat disappointing: a negligible increase in TCP has been achieved, but

at the cost of a very large increase in NTCP. However this is not really unexpected: GBM’s

are extremely di�cult to treat and one of several reasons is the low radiosensitivity of some

13

Dose-volume histograms (DVH) for the whole volume of the simulated head (left panel) and for the planning target volume (right panel). DVH’s are empirical cumulative distributions of dose that are often used in radiotherapy, but they are read off differently from usual cumulative distributions. For instance, from the histogram on the right we find that about 70% of all voxels receives a dose larger than 60 Gy, and that about 30% of all voxels receives a dose larger than 100 Gy.

Dose-volume histograms


Simulation with 4 beams


Simulation with 3 beams and doubled beam width


By carefully adjusting the beam parameters we can optimize the results of radiation therapy.

This simple example shows how to use the basic principles, however:

• example limited to 2D (real treatment plans must be 3D)• no real physics (intensity does not change because of absorption, no

Compton scattering of photons, etc.)• quantification of damage with simplified TCP and NTCP curves• simple structure with circular symmetry (real cases are much more complex)• no organ-at-risk in the vicinity• ...

logo


Radiation Dose

Dose is the measure of energydeposited in medium by ionizingradiation per unit mass

gray (Gy) = JouleKilogram

Dose deposited in patient ismeasured using a fine cubical grid

cubes are called voxels

8 / 39


Conclusions


Overview

Molecular Biology for the Radiation Oncologist: the 5Rsof Radiobiology meet the Hallmarks of Cancer

K. Harrington*y, P. Jankowska*, M. Hingoraniy*The Institute of Cancer Research, Targeted Therapy Laboratory, Cancer Research UK,

Centre for Cell and Molecular Biology, London SW3 6JB, UK;yHead and Neck Unit, Royal Marsden Hospital, London SW3 6JJ, UK

ABSTRACT:Recent advances in our understanding of the biology of cancer have provided enormous opportunities for thedevelopment of novel therapies against specific molecular targets. It is likely that most of these targeted therapies willhave only modest single agent activities but may have the potential to accentuate the therapeutic effects of ionisingradiation. In this introductory review, the 5Rs of classical radiobiology are interpreted in terms of their relationshipto the hallmarks of cancer. Future articles will focus on the specific hallmarks of cancer and will highlight theopportunities that exist for designing new combination treatment regimens. Harrington, K. et al. (2007). ClinicalOncology 19, 561e571

ª 2007 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved.

Key words: Hallmarks of cancer, molecular biology, radiobiology, targeted therapy

Introduction

It is an extremely exciting time to be a clinical oncologist.We work in a specialty that has undergone seismic shifts inroutine practice in the last decade. These include: (i) theapplication of technological advances in radiation delivery;(ii) the demonstration of the superiority of chemoradio-therapy over radiotherapy alone for a range of tumour types;and (iii) the development of novel targeted therapies forintegration within standard combination strategies. Indeed,it is not an exaggeration to claim that clinical oncology iscurrently in the middle of a renaissance and at its heartis the realisation that combinations of start-of-the-artradiotherapy, chemotherapy and new targeted drugs willprobably yield significant therapeutic advantages toa large number of patients in the foreseeable future.

After decades of stagnation, technological developmentshave brought three-dimensional conformal radiotherapyand intensity-modulated radiotherapy within the reach ofmost departments [1]. In fact, with a few exceptions, thepace of introduction of the new technologies has out-stripped our ability (or willingness) to conduct carefullycontrolled randomised clinical trials comparing them withconventional radiotherapy. In addition, the use of partic-ulate radiation (protons, carbon ions) is receiving renewedattention and large collaborative projects have beenestablished in Europe and the USA [2,3]. The next 20 yearswill probably require significant research effort by clinicaloncologists as they focus on implementing the newtechnologies for radiation delivery.

More recently, after much previous debate and contro-versy, meta-analyses have clearly shown the clinical benefitof adding concomitant cytotoxic chemotherapy to radio-therapy in a number of tumour types in both radical andadjuvant postoperative settings [4e9]. As a consequence,concomitant chemoradiotherapy has become the standardof care for many tumour types. This change in practice hasbrought with it new problems, including the selection ofappropriate patients for chemoradiotherapy and themanagement of the increased acute (and possibly late)toxicity of chemoradiotherapy [10,11].

While these changes in clinical practice have been takingplace, we have witnessed fundamental changes in ourunderstanding of the biology of cancer and, as a conse-quence, we are just beginning to reap the rewards of thisresearch in the form of novel targeted agents. For example,a recent phase III randomised study of radiation with orwithout a targeted monoclonal antibody (cetuximab) inpatients with head and neck cancer showed a verysignificant advantage for the combined regimen [12] andthis agent is now undergoing evaluation in randomisedstudies with chemoradiotherapy [13]. Undoubtedly, thenext decade will see a wide range of new targeted drugscoming to the clinic for use alongside standard chemo-radiotherapy regimens. Indeed, it is not inconceivable thatin due course some of these agents may replace cytotoxicchemotherapy in combination strategies.

Therefore, in addition to possessing expertise with thenew technologies, clinical oncologists will be expected toconduct and assess trials of novel targeted agents in

0936-6555/07/190561þ11 $35.00/0 ª 2007 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved.

Clinical Oncology (2007) 19: 561e571doi:10.1016/j.clon.2007.04.009


Overview







Introduction








Overview







Introduction








Overview







Introduction








(from Harrington &al.: “Molecular Biology for the Radiation Oncologist: the 5Rs of Radiobiology meet the Hallmarks of Cancer”, Cell 144 (2011) 646)


Weinberg [14] provided a useful framework for thinkingabout the steps that play a key part in the development,progression, spread and response to treatment of mostcancers (Fig. 1). From analyses of the evolution of tumoursthrough their pre-malignant precursors to their franklymalignant metastatic manifestations, two things havebecome clear: (i) single genetic abnormalities are rarelysufficient to cause cancer; and (ii) the sequence in whichmultiple abnormalities accumulate is not necessarilyimportant. Nonetheless, each of the steps in the processof malignant transformation represents an opportunity fortherapeutic intervention and many of them have specificrelevance to the practice of radiation oncology. As we shall

see, many of the hallmarks of cancer can be invoked (singlyor in combination) to explain the fundamental observationsenshrined in the 5Rs of classical radiobiology (Fig. 2).

Self-sufficiency in Growth Factors

A general scheme for the role of growth factor receptorsand their ligands in promoting cell growth (and otherfunctions) is shown in Fig. 3. Binding of a growth factor (thecognate ligand) to its specific ligand-binding domain on theextracellular component of the receptor leads to a signalbeing passed from the membrane to the nucleus viaa cascade of intermediary messengers such that the bindingof a protein on the cell surface is able to influence thebehaviour of the cell [25,26].

Under normal circumstances, the activation of growthfactor receptors is very tightly controlled d as is thesynthesis and release of the ligands that stimulate them.Cancer cells frequently usurp normal signalling throughgrowth factor receptors and use this to promote un-restrained cell division. Cancer cells exploit three mainstrategies for achieving autonomy in growth factors: (i)they manufacture and release their own growth factorsthat are able to stimulate their own receptors (autocrinesignalling) and those of their immediate neighbours (para-crine signalling) [27,28]; (ii) they alter the number,structure or function of the growth factor receptors ontheir surface such that they are more likely to send a growthsignal to the nucleus (even in the absence of the cognateligand) [29,30]; (iii) they deregulate the growth signallingpathway downstream of the growth factor receptor such

REPAIR

REPOPULATION

REDISTRIBUTION

REOXYGENATION

RADIOSENSITIVITY

GROWTH FACTORSELF-SUFFICIENCY

INSENSITIVITY TO ANTI-GROWTHFACTOR SIGNALLING

EVASION OF APOPTOSIS

ANGIOGENESIS

IMMORTALISATION BY TELOMERASE REACTIVATION

INVASION AND METASTASIS

Fig. 2 e Potential relationships between the 5Rs of radiobiologyand the hallmarks of cancer.

AngiogenesisInvasion

Proliferation

Gene transcriptionAnti-apoptosis

DNA repair

1. Ligand binds to extracellular domain

2. Intracellular domain undergoes modification

3. Activation of signal transduction pathways

-PP-

Fig. 3 e Simplified diagram illustrating activation of a tyrosine kinase growth factor receptor. Binding of ligand to the extracellular domain ofthe receptor leads to dimerisation, phosphorylation of the intracellular domain and signal transduction through second messengers that leadto phenotypic changes.

563MOLECULAR BIOLOGY FOR THE RADIATION ONCOLOGIST

number and thus maintenance of normal tissue architecture andfunction. Cancer cells, by deregulating these signals, becomemasters of their own destinies. The enabling signals areconveyed in large part by growth factors that bind cell-surfacereceptors, typically containing intracellular tyrosine kinasedomains. The latter proceed to emit signals via branched intra-cellular signaling pathways that regulate progression throughthe cell cycle as well as cell growth (that is, increases in cellsize); often these signals influence yet other cell-biological prop-erties, such as cell survival and energy metabolism.Remarkably, the precise identities and sources of the prolifer-

ative signals operating within normal tissues were poorly under-stood a decade ago and in general remain so. Moreover, we stillknow relatively little about the mechanisms controlling therelease of these mitogenic signals. In part, the understandingof these mechanisms is complicated by the fact that the growthfactor signals controlling cell number and position within tissuesare thought to be transmitted in a temporally and spatially regu-lated fashion from one cell to its neighbors; such paracrinesignaling is difficult to access experimentally. In addition, thebioavailability of growth factors is regulated by sequestration inthe pericellular space and extracellular matrix, and by the actionsof a complex network of proteases, sulfatases, and possiblyother enzymes that liberate and activate them, apparently ina highly specific and localized fashion.The mitogenic signaling in cancer cells is, in contrast, better

understood (Lemmon and Schlessinger, 2010; Witsch et al.,2010; Hynes and MacDonald, 2009; Perona, 2006). Cancer cellscan acquire the capability to sustain proliferative signaling ina number of alternative ways: They may produce growth factorligands themselves, to which they can respond via the expres-sion of cognate receptors, resulting in autocrine proliferativestimulation. Alternatively, cancer cells may send signals to stim-ulate normal cells within the supporting tumor-associatedstroma, which reciprocate by supplying the cancer cells withvarious growth factors (Cheng et al., 2008; Bhowmick et al.,2004). Receptor signaling can also be deregulated by elevatingthe levels of receptor proteins displayed at the cancer cell

Figure 1. The Hallmarks of CancerThis illustration encompasses the six hallmarkcapabilities originally proposed in our 2000 per-spective. The past decade has witnessedremarkable progress toward understanding themechanistic underpinnings of each hallmark.

surface, rendering such cells hyperre-sponsive to otherwise-limiting amountsof growth factor ligand; the sameoutcome can result from structural alter-ations in the receptor molecules thatfacilitate ligand-independent firing.Growth factor independence may also

derive from the constitutive activation ofcomponents of signaling pathways oper-ating downstream of these receptors,obviating the need to stimulate thesepathways by ligand-mediated receptor

activation. Given that a number of distinct downstream signalingpathways radiate from a ligand-stimulated receptor, the activa-tion of one or another of these downstream pathways, forexample, the one responding to the Ras signal transducer,may only recapitulate a subset of the regulatory instructionstransmitted by an activated receptor.Somatic Mutations Activate Additional DownstreamPathwaysHigh-throughput DNA sequencing analyses of cancer cellgenomes have revealed somatic mutations in certain humantumors that predict constitutive activation of signaling circuitsusually triggered by activated growth factor receptors. Thus,we now know that !40% of human melanomas containactivating mutations affecting the structure of the B-Raf protein,resulting in constitutive signaling through the Raf to mitogen-activated protein (MAP)-kinase pathway (Davies and Samuels2010). Similarly, mutations in the catalytic subunit of phosphoi-nositide 3-kinase (PI3-kinase) isoforms are being detected inan array of tumor types, which serve to hyperactivate the PI3-kinase signaling circuitry, including its key Akt/PKB signaltransducer (Jiang and Liu, 2009; Yuan and Cantley, 2008). Theadvantages to tumor cells of activating upstream (receptor)versus downstream (transducer) signaling remain obscure, asdoes the functional impact of crosstalk between the multiplepathways radiating from growth factor receptors.Disruptions of Negative-Feedback Mechanisms thatAttenuate Proliferative SignalingRecent results have highlighted the importance of negative-feedback loops that normally operate to dampen various typesof signaling and thereby ensure homeostatic regulation of theflux of signals coursing through the intracellular circuitry (Wertzand Dixit, 2010; Cabrita and Christofori, 2008; Amit et al.,2007; Mosesson et al., 2008). Defects in these feedback mech-anisms are capable of enhancing proliferative signaling. Theprototype of this type of regulation involves the Ras oncoprotein:the oncogenic effects of Ras do not result from a hyperactivationof its signaling powers; instead, the oncogenic mutationsaffecting ras genes compromise Ras GTPase activity, which

Cell 144, March 4, 2011 ª2011 Elsevier Inc. 647

number and thus maintenance of normal tissue architecture andfunction. Cancer cells, by deregulating these signals, becomemasters of their own destinies. The enabling signals areconveyed in large part by growth factors that bind cell-surfacereceptors, typically containing intracellular tyrosine kinasedomains. The latter proceed to emit signals via branched intra-cellular signaling pathways that regulate progression throughthe cell cycle as well as cell growth (that is, increases in cellsize); often these signals influence yet other cell-biological prop-erties, such as cell survival and energy metabolism.Remarkably, the precise identities and sources of the prolifer-

ative signals operating within normal tissues were poorly under-stood a decade ago and in general remain so. Moreover, we stillknow relatively little about the mechanisms controlling therelease of these mitogenic signals. In part, the understandingof these mechanisms is complicated by the fact that the growthfactor signals controlling cell number and position within tissuesare thought to be transmitted in a temporally and spatially regu-lated fashion from one cell to its neighbors; such paracrinesignaling is difficult to access experimentally. In addition, thebioavailability of growth factors is regulated by sequestration inthe pericellular space and extracellular matrix, and by the actionsof a complex network of proteases, sulfatases, and possiblyother enzymes that liberate and activate them, apparently ina highly specific and localized fashion.The mitogenic signaling in cancer cells is, in contrast, better

understood (Lemmon and Schlessinger, 2010; Witsch et al.,2010; Hynes and MacDonald, 2009; Perona, 2006). Cancer cellscan acquire the capability to sustain proliferative signaling ina number of alternative ways: They may produce growth factorligands themselves, to which they can respond via the expres-sion of cognate receptors, resulting in autocrine proliferativestimulation. Alternatively, cancer cells may send signals to stim-ulate normal cells within the supporting tumor-associatedstroma, which reciprocate by supplying the cancer cells withvarious growth factors (Cheng et al., 2008; Bhowmick et al.,2004). Receptor signaling can also be deregulated by elevatingthe levels of receptor proteins displayed at the cancer cell

Figure 1. The Hallmarks of CancerThis illustration encompasses the six hallmarkcapabilities originally proposed in our 2000 per-spective. The past decade has witnessedremarkable progress toward understanding themechanistic underpinnings of each hallmark.

surface, rendering such cells hyperre-sponsive to otherwise-limiting amountsof growth factor ligand; the sameoutcome can result from structural alter-ations in the receptor molecules thatfacilitate ligand-independent firing.Growth factor independence may also

derive from the constitutive activation ofcomponents of signaling pathways oper-ating downstream of these receptors,obviating the need to stimulate thesepathways by ligand-mediated receptor

activation. Given that a number of distinct downstream signalingpathways radiate from a ligand-stimulated receptor, the activa-tion of one or another of these downstream pathways, forexample, the one responding to the Ras signal transducer,may only recapitulate a subset of the regulatory instructionstransmitted by an activated receptor.Somatic Mutations Activate Additional DownstreamPathwaysHigh-throughput DNA sequencing analyses of cancer cellgenomes have revealed somatic mutations in certain humantumors that predict constitutive activation of signaling circuitsusually triggered by activated growth factor receptors. Thus,we now know that !40% of human melanomas containactivating mutations affecting the structure of the B-Raf protein,resulting in constitutive signaling through the Raf to mitogen-activated protein (MAP)-kinase pathway (Davies and Samuels2010). Similarly, mutations in the catalytic subunit of phosphoi-nositide 3-kinase (PI3-kinase) isoforms are being detected inan array of tumor types, which serve to hyperactivate the PI3-kinase signaling circuitry, including its key Akt/PKB signaltransducer (Jiang and Liu, 2009; Yuan and Cantley, 2008). Theadvantages to tumor cells of activating upstream (receptor)versus downstream (transducer) signaling remain obscure, asdoes the functional impact of crosstalk between the multiplepathways radiating from growth factor receptors.Disruptions of Negative-Feedback Mechanisms thatAttenuate Proliferative SignalingRecent results have highlighted the importance of negative-feedback loops that normally operate to dampen various typesof signaling and thereby ensure homeostatic regulation of theflux of signals coursing through the intracellular circuitry (Wertzand Dixit, 2010; Cabrita and Christofori, 2008; Amit et al.,2007; Mosesson et al., 2008). Defects in these feedback mech-anisms are capable of enhancing proliferative signaling. Theprototype of this type of regulation involves the Ras oncoprotein:the oncogenic effects of Ras do not result from a hyperactivationof its signaling powers; instead, the oncogenic mutationsaffecting ras genes compromise Ras GTPase activity, which

Cell 144, March 4, 2011 ª2011 Elsevier Inc. 647

(from Hanahan & Weinberg: “Hallmarks of Cancer: The Next Generation”, Cell 144 (2011) 646)



initial impression of the efficacy of the new combination incomparison with the standard therapy. In theory, if thepre-clinical data are able to define suitable groups forclinical trials on the basis of the underlying biology of thecancer, these phase II studies may have sufficientstatistical power to yield answers that previously haverequired phase III trials. As a result, it is probable thatphase III studies may evolve to require smaller numbers ofpatients with tumours that are biologically homogeneousin which testing of a new targeted drug representsa rational strategy. Such careful target definition in earlyphase IeII studies will hopefully avoid the sort of debaclethat occurred with the testing of gefitinib in combinationwith gemcitabine and cisplatin (INTACT 1) or paclitaxeland carboplatin (INTACT 2) in two large phase III studiesinvolving more than 2000 patients before the importanceof EGFR mutation status to the likelihood of response wasunderstood [89e91].

Development of Targeted TherapiesBased on an Integration of the 5Rs ofRadiobiology and the Hallmarks of Cancer

The real challenge that faces clinical oncologists at thistime is the question of selecting the best candidatemolecules for evaluation alongside radical chemoradiother-apy. The obvious corollary of this challenge (if we acceptthe premise that we will not be able to add cocktails ofthree, four or five new drugs concomitantly to standardchemoradiotherapy regimens) is that we will need toexploit the potential opportunities of other targeted agentsin the neoadjuvant or adjuvant setting. Thus, we will needto use an integrated understanding of the biology of cancerderived from the 5Rs of radiobiology and the hallmarks ofcancer to guide treatment design. Figure 6 providesa schema that illustrates the conceptual links betweenthe 5Rs of radiobiology and the hallmarks of cancer.Consideration of these links allows us to build up a pictureof a future trial design that aims to hit the various targetson offer at a biologically optimal time (these timings may

differ for different tumour types). By way of illustration,we shall consider how we may build a strategy forintroducing new drugs into a standard chemoradiotherapyapproach in head and neck cancer (Fig. 6).

The link between growth factor self-sufficiency and all 5Rsof radiobiology represents a compelling case for theconcomitant use of drugs that interact with this processduring chemoradiotherapy. Indeed, the benefit of using EGFRblockade with radiotherapy has been shown in a randomisedsetting in patients with head and neck cancer [12]. Theimportance of the restoration of sensitivity to anti-growthsignals to tumour cell repopulation, redistribution andradiosensitivity means that drugs that are able to targetthese pathways should also be considered for concomitantuse with chemoradiotherapy. Similarly, the link between theevasion of apoptosis and DNA damage repair and radiosen-sitivity argues strongly for the administration of drugs thatre-set the apoptotic balance of cancer cells in a pro-apoptotic direction concurrently with radiation. Therefore,we immediately have three groups of drugs competing fora place in a combination regimen with chemoradiotherapy. Inorder to exploit each of these opportunities to the maximumextent, it might be reasonable to deliver a drug thatmodulates tumour apoptosis throughout the whole treat-ment course, but schedule an EGFR blocker for the first 4weeks of treatment and a drug that restores anti-growthfactor signalling for the last 3 weeks of treatment (or viceversa). Such a strategy may represent a rational approach totargeting accelerated repopulation of tumour clonogens inthe latter half of treatment.

The role of angiogenesis in establishing and maintaininga tumour blood supply makes this an ideal candidate fortargeting in a neoadjuvant or adjuvant setting. In additionto having anti-tumour effects, anti-angiogenic agents maybe capable of the ‘normalisation’ of aberrant tumour-associated blood vessels with a resulting improvementin tumour oxygenation and drug delivery [63]. There-fore, short duration neoadjuvant administration of anti-angiogenic drugs may prime tumours for a better responseto subsequent chemoradiation. The growth dependence ofmetastatic colonies on the activation of angiogenesis alsomeans that anti-angiogenic drugs may be very goodcandidates for adjuvant use against micrometastatic dis-ease. The role of telomerase reactivation as an overarchingfeature of cancer biology means that targeting this hallmarkmay yield therapeutic gains at any time during thetreatment of cancer. As such, at present, there is nocompelling reason for scheduling agents that target thisprocess concomitantly with chemoradiation. Instead, it maybe advantageous to use anti-telomerase strategies in theneoadjuvant or adjuvant settings.

Finally, it is difficult to define a clear role for agents thattarget tumour cell invasion and metastasis. Nonetheless, itis reasonable to hypothesise that any drugs capable oftargeting this process might have a useful role in neo-adjuvant treatment (before surgery or definitive chemo-radiotherapy) as a means of limiting the local extent ofdisease (anti-invasion) or the likelihood of tumourcell spread (anti-metastasis). Similarly, as these signalling

Induction/Neoadjuvant

Concomitant Adjuvant

Growth factor blockers

Pro-apoptotic drugs

Anti-angiogenic agents

Anti-growth signalling

Anti-telomerase drugs

Anti-invasion/metastasis drugs

Fig. 6

568 CLINICAL ONCOLOGY

A picture of a future trial design that aims to hit the various targets on offer at a biologically optimal time in the case of Head and Neck cancer

(from

Har

ringt

on &

al.:

“Mol

ecul

ar B

iolo

gy fo

r the

Rad

iatio

n On

colo

gist

: th

e 5R

s of R

adio

biol

ogy

mee

t the

Hal

lmar

ks o

f Can

cer”

, Cel

l 144

(201

1)

646)


Overview







Introduction








combination with (chemo)radiotherapy. It is of paramountimportance that the specialty embraces this challenge inorder to ensure that the direction of clinical studies isinformed by sound radiobiological principles, such that thefocus is on maximising the effect of the most importantcomponent of the treatment (i.e. radiation). Failure to riseto this challenge means that clinical oncologists will takea passive role in the development of new strategies and willrun the risk of being relegated to the role of radiationtechnicians.

In this series of review articles, we will discuss the keyadvances in the molecular biology of cancer as theyrelate specifically to the practice of clinical oncology. Inthis introduction we will briefly describe the key featuresor ‘hallmarks’ of cancer [14] as a prelude to subsequentarticles that will explore the potential effect of each ofthese processes on the field of clinical oncology. We willalso attempt to show how information derived fromstudies of the molecular biology of cancer can be used tobreathe new life into the 5Rs of radiobiology and makethem relevant to the new generation of radiationoncologists.

Radiobiological Determinantsof Treatment Outcome

Radiotherapy is an extremely effective treatment forcancer, especially when the disease presents at an earlystage. However, despite its undoubted activity, localisedradiotherapy (with or without cytotoxic chemotherapy)frequently fails to eradicate all of the clonogenic cellswithin a cancer and the tangible reality of this failure isa local or regional recurrence of disease. Alternatively,radiotherapy (with or without cytotoxic chemotherapy)may be used as an intensive local therapy for a disease thathas already slipped the leash and spread to distant sites,with the inevitable consequence of disease recurrenceoutside the radiation portals.

The 4Rs of radiobiology were initially described in anattempt to provide a means of understanding the success orfailure of localised radiotherapy [15]. The differentialrepair of tumour and normal cells between treatmentfractions, the redistribution of cells into more or lessradiosensitive phases of the cell cycle, the repopulation oftumour cells between fractions and the re-oxygenation oftumour cells during treatment were all invoked to explainthe net outcome of radiotherapy. Later, the system wasrevised to include intrinsic radiosensitivity in the 5Rs ofradiobiology [16]. With a few exceptions, this final additionto the quintet was an admission of our inability to explain atthe mechanistic level the different radioresponsiveness ofdiseases like seminoma, lymphoma, glioma and melanoma.

Nonetheless, the 5Rs have served an extremely impor-tant function in providing a framework within which toexamine new therapeutic strategies from the point of viewof both tumour and normal cells. Each of the Rs can beviewed as a double-edged sword such that changes canoccur in either direction to increase or decrease the net

therapeutic effect. For example, if a tumour cell hasacquired a defect in its DNA repair pathway, it is more likelythan an adjacent normal cell to be killed by a dose ofradiation [17e19]. However, the abnormal DNA repairpathway may already have allowed the tumour cell toaccumulate non-lethal mutations in other important genesthat allow it to tolerate unrepaired DNA damage (or torepair it in an inaccurate manner that only serves toenhance genetic instability). Similarly, the enhancedtumour cell division that occurs during a course ofradiotherapy is generally viewed negatively as the drivingforce behind accelerated repopulation, but it may alsomake a tumour cell more susceptible to radiation-induceddeath by causing it to enter mitosis with unrepaired DNAdamage (so-called mitotic catastrophe).

Our new insights into the molecular biology of cancerhave now put us in a position to reinterpret the classical 5Rsof radiobiology in terms of their underlying mechanisms. Aswe shall see below, a direct one-to-one translation of eachof the Rs of radiobiology into a single biological mechanismis not possible. For example, DNA damage repair can beinfluenced by growth factor receptor autonomy and evasionof apoptosis. However, the particular strength of describingcancer in terms of its molecular biological hallmarks is thatit leads naturally into a discussion of potential newtargeted therapies that may favourably modulate thetumour response and increase the therapeutic index [20].

The Molecular Biological Hallmarksof Cancer

There is ample evidence to support the hypothesis thathuman tumours arise as part of a sequential multi-stepprocess, with each step reflecting the accumulation ofgenetic alterations that confer a survival advantage on theevolving malignant cell population [21e24]. Hanahan and

MalignantPhenotype

GrowthFactor

Independence

Evasion of Apoptosis

Insensitivity toAnti-growth Signals

SustainedAngiogenesis

Immortalisation by Reactivation of

Telomerase

Tissue Invasionand Metastasis

Fig. 1 e The six hallmarks of cancer.

562 CLINICAL ONCOLOGY


introduction to radiobiology lesson 6 - …milotti/didattica/introductory... · edoardo milotti -...

Documents