18 FDG PET/CT in the diagnosis of Malignant Peripheral Nerve

Sheath Tumours

VS Warbey, RE Ferner, JT Dunn, E Calonje, MJ O’Doherty

St Thomas’ Clinical PET Centre and Department of Clinical Neurosciences

Guy’s, King’s and St Thomas’ School of Medicine

Introduction

MPNST in NF1 – a diagnostic challenge

Overlap of clinical manifestations

MRI identifies site and extent, not reliable in detecting malignant change

Histology is the final arbiter of malignancy

Management requires a Specialist MDM approach

Background

Previous work by our group and others has concluded that 18FDG-PET is helpful in determining malignant change in neurofibromas

Significant difference in SUVmax between benign and malignant lesions with delayed imaging at approximately 200 minutes

SUVmax on delayed imaging < 2.5 2.5 - 3.5 > 3.5

Ferner et al. J Neurol Neurosurg Psychiatry 2000;68:353-357

Bredella et al.AJR 2007;189:928-935

Ferner et al. Ann Oncol 2008; 19(2):390-4

Aims

To evaluate the sensitivity of PET/CT

To clarify the value of early and delayed imaging

To re-validate the current cut-off values for identification of malignant change within neurofibromas using PET/CT

To examine the relationship between SUV and tumour grade

Methods

Patients with symptomatic neurofibromas referred for 18FDG PET/CT were identified from the reports archive

Early and delayed imaging (90 minutes, 4 hours)

SUVmax measured

Classified as malignant: SUVmax rose to > 3.5 on delayed imaging

Histological correlation

Benign

Early Imaging

Late Imaging

Malignant (1)

Early Imaging

Late Imaging

Malignant (2)

Early Imaging

Late Imaging

Results (1)

97 studies were identified from the PET/CT reports archive over a 35-month period from August 2004 to April 2008

Final analysis: 69 studies in 62 patients, 85 neurofibromas Exclusions: No delayed imaging/no focal uptake, alternative

diagnosis, repeat lesion 31 males, 31 females Mean age 31 years, age range 9 – 86

Median imaging times Early 101 minutes (1 hour 41 minutes) Late 252 minutes (4 hours 12 minutes)

Results (2) On the basis of semi-quantitative analysis PET/CT classified:

43 malignant 42 benign

Type of Tumour Negative on 18FDG PET/CT

Positive on 18FDG PET/CT

Neurofibroma 2 6

Atypical 1 9

Low Grade 0 11

High Grade 0 10

Sensitivity 0.97 (95% CI 0.81-0.99)Specificity 0.87 (95% CI 0.74-0.95)

Results (3) - SUVmax Comparsion

Mean SUVmax Benign Lesions on PET

Malignant Lesions on PET

Early 2.0

(CI 1.8-2.3)

7.0

(CI 5.6-8.4)

Late 1.9

(CI 1.7-2.2)

8.1

(CI 6.5-9.6)

Early vs. delayed imaging ([type irrelevant] F 1,83 = 9.98, p=0.0022)Benign vs. malignant ([time irrelevant] F 1,83 = 56.14, p<<0.0001)

Significant interaction effect between time and tumour type (F 1,83 = 14.72, p=0.00017)

SUVmax early vs. delayed imaging for tumours classified as malignant on PET/CT (p=0.0005)

Results (4) - Histology vs SUVmax

Mean SUVmax

Atypical Low Grade High Grade

Early 5.1 7.3 12.0

Late 5.6 7.8 13.7

SUVmax between tumour types ([time irrelevant] F2,27 = 7.91, p=0.002)

SUVmax early vs. delayed imaging (F1,27= 10.58, p=0.003)

Results (5) - Histology vs SUVmax

0

5

10

15

20

25

30

Benign Atypical Low-Grade High-Grade

SU

Vm

ax

Conclusions

Recommend early and delayed PET/CT imaging for accurate lesion characterisation

Continue to recommend a cut-off value SUVmaxD=3.5

Acceptable to maintain a high sensitivity at expense of false positive studies

First study that has demonstrated a correlation between SUVmax and tumour grade