GASTROINTESTINAL STROMAL TUMORS (GIST) IN NEUROFIBROMATOSIS 1 (NF1) PATIENTS: A CLINICOPATHOLOGIC ANALYSIS OF NINE CASES Elena Fumagalli, Paola Coco, Elena Palassini, Palma Dileo,  Rossella Bertulli, Paolo G. Casali Istituto Nazionale Tumori Milan, Italy

This series 5 females : 4 males Median age 51 yrs (range 36-59) Localized disease in all cases, with multiple lesions in 6 pts Site: stomach and duodenum 1pt duodenum 2 pts jejunum 1 pt ileum 4 pts other 1 pt Risk stratification: high 2 pts intermediate 1 pt low/very low 6 pts

This series: pts with multifocal disease (5 low risk, 1 high risk) 2 pts 4 pts

Multifocal disease Study Patients Multifocality Yantiss et al. Mod Pathol 2005 3 100% Takazawa et al. Am J Surg Pathol 2005 8 90% Mussi et al. Clin Cancer Res 2008 28 43% Maertens et al. Hum Mol Genet 2006 Andersson et al. Am J Surg Pathol 2005 11 91% Miettinen et al. Am J Surg Pathol 2006 45 62% Kinoshita et al. J Pathol 2004 7 Kramer et al. World J Gastroenterol 2007 1 Kang et al. Am J Surg Pathol 2007 5

“The presence of multiple GIST in NF1 might reflect a distinct rate-limiting step in oncogenesis compared with sporadic GIST. While a broad spectrum of inactivating genetic mechanism might lead to suppression of the wild type NF1 allele and GIST formation in NF1 pts, only a limited set of specific activating mutations in KIT/PDGFRA will result in sporadic GIST”

This series: molecular genetics 8 pts: WT 1 pt (gastric + duodenal GIST): PDGFRA exon 18 (D842V) mutation in the gastric GIST; WT in the duodenal GIST

affecting the A-loop of PDGFRA Immunohistochemistry D842V in exon 18 of PDGFRA G A G W C A T C A PDGFRA D842V affecting the A-loop of PDGFRA

KIT & PDGFRA mutations Study Patients Mutations Takazawa et al. Am J Surg Pathol. 2005 9 KIT L558L, P627L ex11, I653T ex13; PDGFRA P589S ex12, R822S Yantiss et al. Mod Pathol. 2005 3 KIT V559D ex11 in 3 tumors →1 pt Cheng et al. Dig Dis Sci. 2004 KIT delWK557-558 ex11 in 1 pt Mussi et al. Clin Cancer Res.2008 25 KIT delV560 ex11,dupl A502_Y503 ex9; PDGFRA D842V ex18 Maertens et al. Hum Mol Genet. 2006 None Lee et al. Dig Dis Sci. 2006 1 Andersson et al. Am J Surg Pathol. 2005 12 Miettinen et al. Am J Surg Pathol. 2006 15 Steward et al. J Med Genet. 2007 2 Kinoshita et al. J Pathol. 2004 7 Kang et al. Am J Surg Pathol. 2007 5

KIT/PDGFRA KIT/PDGFRA

This series: pts with NET 1 pt: pheochromocytoma + ampullary neuroendrocrine tumor pheocromocitoma neuroendocrine tumour

GIST and NET in NF1 Study Patients NET Karatzas et al. Eur J Surg Oncol, 2000 1 Somatostatinoma Andersson et al. Am J Surg Pathol, 2005 4 Pheocromocytoma Usui et al. J Gastroenterol,2002 Kramer et al. Z Gastroenterol, 2005 Duodenum neuroendocrine carcinoma Lisewski et al. Int Semin Surg Oncol, 2006 Bilateral pheocromocytoma Kramer et al. World J Gastroenterol, 2007 Erem et al. J Endocrinol Investigation, 2007 Bumming et al. Scand J Gastroenterol, 2006 Mussi et al. Clin Cancer Res, 2008 2 Pheocromocytoma Somatostatinoma Juergens et al. AJR, 2006 Duodenal somatostatinoma Suzuki et al. J Gastroenterol, 2004 Masanobu et al. J Gastroenterol, 2002

SUV 5

SUV 4.5 SUV 4.7

SUV 3.8

PET in NF1 Study Patients FDG uptake (SUV max) Ferner et al. Ann Oncol, 2008 105 (80 PN, 29 MPNST, 5 AN) Range 3.9-7.7; 1.3-1.8 Fisher et al. J Neurooncol, 2008 13 pts (19 PN) Range 0.9-4 (median 1.5) Brinkman et al. W J Sur, 2007 1 pt (20 PN) No correlation Bredella et al. AJR, 2007 45 pts (24 MPNST, 26 PN) Range 8.5+/-0.63; 1.5+/-0.37 Brenner et al. Eur J Nucl Med Mol Imaging, 2006 16 pts SUV<3; SUV>3 Correlation with outcome Cardona et al. Eur J Sur Oncol, 2003 (12 PN, 13 MPNST) Range 1.8-12.3; 0.5-1.8 Solomon et al. Clin Nucl Med, 2001 Ferner et al. J Neurol Neurosurg Psychiatry, 2000 18 pts (23 PN) Range 2.7-8.4; 0.56-3.3 overlap 2.7-3.3

FDG PET distinguishes MPNSTs from benign neurogenic tumours with 100% sensivity and 83% specificity at an SUV cut-off value = 1.8 SUV level predicts outcome in NF1 pts with MPNSTs SUV cut-off value = 3 No correlation with histological grading

FDG PET and PET TC is Significant difference between mean SUV of malignant and benign tumours Overlap with SUVmax = 2.5-3.5  lesions should be reviewed clinically Time for measuring SUV 240 min FDG PET and PET TC is sensitive and specific for MPNST in NF1 pts Research with different tracers to predict tumour grade

This series: prognosis 8 pts: alive and well median follow-up = 23 months (range = 6-97 months) 1 pt: metastatic disease (baseline: >5 cm; >10/50 HPF)  death after progression to Imatinib

GIST in NF1: prognosis Author Patients Median follow-up Miettinen et al. Am J Surg Pathol 2006 45 Median FU 13.6 yrs 20 pts alive Mussi et al. Clin Cancer Res 2008 28 5 yrs disease specif survival 54.3% (median not reached); event free survival 43.9% (median 48 mos) Median FU (metastaic GIST): 33 mos Maertens et al. Hum Mol Genet 2006 3 No metastases, 1 pt died due to complications after surgery Median FU 2 yrs Yantiss et al. Mod Pathol 2005 1 pt died due to metastases: KIT V559D ex11 in 3 tumors Andersson et al. Am J Surg Pathol 2005 15 Median FU 3 yrs 6 pts alive

This series: 1 pt treated preoperatively baseline + 4 w

This series: secondary resistance 1 pt (ileal GIST): Kit exon 17 (D820N) mutation in metastatic lesions; WT in primary tumour D820N cKit ex 17 immunohistochemistry A G A A T r A T T cKIT Mussi et al. Clin Cancer Res 2008:14 July 15

AntiTK activity

Conclusions Only anecdotal cases of GIST have been reported in NF1, though in the face of a 5-10% risk of developing the disease in this syndrome WT are predominant, but KIT/PDGFRA mutations are occasionally present (both to KIT and PDGFRA) NET may be concurrent, in the face of a 1% risk of these diseases in NF1 Prognosis of WT GIST in NF1, although multifocal, is good, but may be worse for KIT/PDGFRA-mutated GIST AntiTK are often ineffective, but responses have been occasionally reported (and secondary resistance may arise) GIST in NF1 are often positive on PET scan, but specificity may be problematic against other lesions, including neurofibromas

elena.fumagalli@istitutotumori.mi.it