Radiology-NLP–derived real-world progression endpointing and clinicogenomic correlates after 177Lu-DOTATATE PRRT in pancreatic neuroendocrine tumors in MSK-CHORD 2024.

A Amro Al-Omari (Department of Medicine, Trinity Health Livonia, Livonia, MI) U Umer Javaid (Department of Mechatronics Engineering, College of Electrical and Mechanical Engineering, NUST 1 , Rawalpindi,) A Alice Nassar A Ayham Al-Omari (Detroit Medical Center - Wayne State University, Detroit, MI) R Rana Uzair Ahmad (8Trinity Health Livonia, Michigan, Livonia, United States) U Ushna Gul (Department of Medicine, Trinity Health Livonia, Livonia, MI)

Abstract

e16329 Background: Predictive biomarkers for peptide receptor radionuclide therapy (PRRT) in pancreatic neuroendocrine tumors (PanNET) are limited, and existing clinicogenomic series are small. MSK-CHORD links tumor sequencing with NLP-derived radiology features and radiology-anchored progression timelines, enabling scalable real-world endpointing. Methods: PanNET were identified by OncoTree (PANET) in MSK-CHORD 2024 (cBioPortal). PRRT exposure was defined by treatment timeline agent “Lutetium Lu-177 dotatate”; index was first PRRT start. Somatic non-silent alterations (MSK-IMPACT) in MEN1, DAXX, and ATRX were evaluated. Radiology-NLP metastatic site indicators were used as baseline phenotypes. Primary endpoint was time to first post-index radiology-annotated progression (PROGRESSION = Y); censoring occurred at the last post-index progression assessment. Kaplan–Meier estimated time to progression (TTP); exploratory Cox proportional hazards models evaluated covariates. Results: Among 259 PanNET, 55 received 177Lu-DOTATATE. Median age was 65 years (IQR 56–72); 55% were male. Baseline NLP-derived metastatic sites included bone 44% and lung 38% (liver 100%). Median prior systemic agents before PRRT was 2 (IQR 1–3). DAXX/ATRX alterations were present in 47% and MEN1 in 53%. Forty-nine (89%) had ≥1 post-PRRT progression assessment; 39 progressed. Median TTP was 14.9 months and 12-month progression-free probability was 59%. DAXX/ATRX-altered tumors showed longer TTP than wild-type (16.6 vs 11.9 months; HR 0.80, 95% CI 0.42–1.52; p = 0.49). Each additional prior systemic agent was associated with shorter TTP (HR 1.16 per agent; 95% CI 1.02–1.32; p = 0.026). Conclusions: A CHORD-based, radiology-derived progression endpoint enables reproducible real-world assessment of PRRT outcomes in PanNET and supports hypothesis generation for clinicogenomic and radiology-phenotype factors associated with post-PRRT progression. PRRT-treated PanNET (n=55), stratified by DAXX/ATRX alteration. Overall DAXX/ATRX altered DAXX/ATRX wild-type n 55 26 29 Age, median (IQR), y 65 (56–72) 68.5 (63–74) 61 (50–68) Male, n (%) 30 (55) 15 (58) 15 (52) Bone mets (NLP), n (%) 24 (44) 12 (46) 12 (41) Lung mets (NLP), n (%) 21 (38) 8 (31) 13 (45) Prior agents, median (IQR) 2 (1–3) 2 (1–3) 2 (2–3) MEN1 altered, n (%) 29 (53) 19 (73) 10 (34) Evaluable for TTP, n 49 23 26 Median TTP, months (KM) 14.9 16.6 11.9 12-mo progression-free, % 59 72 48

Article Details

Volume / Issue Vol. 44, Issue 16_suppl
Published June 01, 2026
ISSN 0732-183X
Publisher Lippincott Williams & Wilkins

Journal Info

Journal of Clinical Oncology

Lippincott Williams & Wilkins

ISSN: 0732-183X Health Sciences

Authors (6)

A

Amro Al-Omari

Department of Medicine, Trinity Health Livonia, Livonia, MI

U

Umer Javaid

Department of Mechatronics Engineering, College of Electrical and Mechanical Engineering, NUST 1 , Rawalpindi,

A

Alice Nassar

A

Ayham Al-Omari

Detroit Medical Center - Wayne State University, Detroit, MI

R

Rana Uzair Ahmad

8Trinity Health Livonia, Michigan, Livonia, United States

U

Ushna Gul

Department of Medicine, Trinity Health Livonia, Livonia, MI