Organoids for Functional Precision Medicine in Advanced Pancreatic Cancer.

BACKGROUND & AIMS: Patient-derived organoids (PDOs) are promising tumor avatars that could enable ex vivo drug tests to personalize patients' treatments in the frame of functional precision oncology. However, clinical evidence remains scarce. This study aims to evaluate whether PDOs can be implemented in clinical practice to benefit patients with advanced refractory pancreatic ductal adenocarcinoma (PDAC). METHODS: During 2021 to 2022, 87 patients were prospectively enrolled in an institutional review board-approved protocol. Inclusion criteria were histologically confirmed PDAC with the tumor site accessible. A panel of 25 approved antitumor therapies (chemogram) was tested and compared to patient responses to assess PDO predictive values and map the drug sensitivity landscape in PDAC. RESULTS: Fifty-four PDOs were generated from 87 pretreated patients (take-on rate, 62%). The main PDO mutations were KRAS (96%), TP53 (88%), and CDKN2A/B (22%), with a 91% concordance rate with their tumor of origin. The mean turnaround time to chemogram was 6.8 weeks. In 91% of cases, ≥1 hit was identified (gemcitabine (n = 20 of 54), docetaxel (n = 18 of 54), and vinorelbine (n = 17 of 54), with a median of 3 hits/patient (range, 0-12). Our cohort included 34 evaluable patients with full clinical follow-up. We report a chemogram sensitivity of 83.3% and specificity of 92.9%. The overall response rate and progression-free survival were higher when patients received a hit treatment as compared to patients who received a nonhit drug (as part of routine management). Finally, we leveraged our PDO collection as a platform for drug validation and combo identification. We tested anti-KRASG12D (MRTX1133), alone or combined, and identified a specific synergy with anti-EGFR therapies in KRASG12D variants. CONCLUSIONS: We report the largest prospective study aiming at implementing PDO-based functional precision oncology and identify very robust predictive values in this clinical setting. In a clinically relevant turnaround time, we identify putative hits for 91% of patients, providing unexpected potential survival benefits in this very aggressive indication. Although this remains to be confirmed in interventional precision oncology trials, PDO collection already provides powerful opportunities for drugs and combinatorial treatment development.

High-resolution and quantitative spatial analysis reveal intra-ductal phenotypic and functional diversification in pancreatic cancer.

A 'classical' and a 'basal-like' subtype of pancreatic cancer have been reported, with differential expression of GATA6 and different dosages of mutant KRAS. We established in situ detection of KRAS point mutations and mRNA panels for the consensus subtypes aiming to project these findings to paraffin-embedded clinical tumour samples for spatial quantitative analysis. We unveiled that, next to inter-patient and intra-patient inter-ductal heterogeneity, intraductal spatial phenotypes exist with anti-correlating expression levels of GATA6 and KRASG12D . The basal-like mRNA panel better captured the basal-like cell states than widely used protein markers. The panels corroborated the co-existence of the classical and basal-like cell states in a single tumour duct with functional diversification, i.e. proliferation and epithelial-to-mesenchymal transition respectively. Mutant KRASG12D detection ascertained an epithelial origin of vimentin-positive cells in the tumour. Uneven spatial distribution of cancer-associated fibroblasts could recreate similar intra-organoid diversification. This extensive heterogeneity with functional cooperation of plastic tumour cells poses extra challenges to therapeutic approaches. © 2023 The Authors. The Journal of Pathology published by John Wiley & Sons Ltd on behalf of The Pathological Society of Great Britain and Ireland.

Molecular deciphering of primary liver neuroendocrine neoplasms confirms their distinct existence with foregut-like profile.

Isolated hepatic localizations of neuroendocrine tumors (NETs) are generally considered as metastatic NETs of unknown primary but could correspond to primary hepatic NETs (PHNETs), a poorly explored entity. We aimed to describe the clinicopathological and molecular features of PHNETs and compare them with other primary NETs. We assembled a retrospective cohort of patients managed for hepatic localization of NET without extra-hepatic primary tumor after exhaustive clinical, imaging, and immunohistochemical characterization. We performed whole-exome sequencing with mutational and copy number analysis. Transcriptomic profiles were compared with pancreatic (n = 31), small-bowel (n = 22), and lung (n = 15) NETs using principal component analysis, unsupervised clustering, and gene set enrichment analysis. Among 27 screened patients, 16 had PHNET (solitary tumor in 63%, median size 11 cm, G2 NETs in 81%) following clinical and pathological review. DNA analyses showed 'foregut-like' genomic profiles with frequent alterations in pathways of Fanconi DNA repair (75%), histone modifiers (58%), adherens junctions (58%), and cell cycle control (50%). The most frequently involved genes were KMT2A (58%), ATM (42%), CDH1, CDKN2C, FANCF, and MEN1 (33% each). Transcriptomic analyses showed that PHNETs clustered closer to foregut (pancreatic, lung) NETs than to midgut (small-bowel) NETs, while remaining a distinct entity with a specific profile. Assessment of potentially predictive biomarkers suggested efficacy of treatments usually active in foregut NETs. In conclusion, PHNETs display a foregut-like molecular profile distinct from other types of NETs, with recurrent molecular alterations. Upon exhaustive work-up to exclude an unrecognized primary tumor, PHNETs should not be considered metastatic NETs from an unknown primary. © 2022 The Pathological Society of Great Britain and Ireland.

Periostin- and podoplanin-positive cancer-associated fibroblast subtypes cooperate to shape the inflamed tumor microenvironment in aggressive pancreatic adenocarcinoma.

Cancer-associated fibroblasts (CAFs) are orchestrators of the pancreatic ductal adenocarcinoma (PDAC) microenvironment. Previously we described four CAF subtypes with specific molecular and functional features. Here, we have refined our CAF subtype signatures using RNAseq and immunostaining with the goal of defining bioinformatically the phenotypic stromal and tumor epithelial states associated with CAF diversity. We used primary CAF cultures grown from patient PDAC tumors, human data sets (in-house and public, including single-cell analyses), genetically engineered mouse PDAC tissues, and patient-derived xenografts (PDX) grown in mice. We found that CAF subtype RNAseq signatures correlated with immunostaining. Tumors rich in periostin-positive CAFs were significantly associated with shorter overall survival of patients. Periostin-positive CAFs were characterized by high proliferation and protein synthesis rates and low α-smooth muscle actin expression and were found in peri-/pre-tumoral areas. They were associated with highly cellular tumors and with macrophage infiltrates. Podoplanin-positive CAFs were associated with immune-related signatures and recruitment of dendritic cells. Importantly, we showed that the combination of periostin-positive CAFs and podoplanin-positive CAFs was associated with specific tumor microenvironment features in terms of stromal abundance and immune cell infiltrates. Podoplanin-positive CAFs identified an inflammatory CAF (iCAF)-like subset, whereas periostin-positive CAFs were not correlated with the published myofibroblastic CAF (myCAF)/iCAF classification. Taken together, these results suggest that a periostin-positive CAF is an early, activated CAF, associated with aggressive tumors, whereas a podoplanin-positive CAF is associated with an immune-related phenotype. These two subpopulations cooperate to define specific tumor microenvironment and patient prognosis and are of putative interest for future therapeutic stratification of patients. © 2022 The Authors. The Journal of Pathology published by John Wiley & Sons Ltd on behalf of The Pathological Society of Great Britain and Ireland.

Whole-Transcriptome Profiling on Small FFPE Samples: Which Sequencing Kit Should Be Used?

RNA sequencing (RNA-Seq) appears as a great tool with huge clinical potential, particularly in oncology. However, sufficient sample size is often a limiting factor and the vast majority of samples from patients with cancer are formalin-fixed paraffin-embedded (FFPE). To date, several sequencing kits are proposed for FFPE samples yet no comparison on low quantities were performed. To select the most reliable, cost-effective, and relevant RNA-Seq approach, we applied five FFPE-compatible kits (based on 3' capture, exome-capture and ribodepletion approaches) using 8 ng to 400 ng of FFPE-derived RNA and compared them to Nanostring on FFPE samples and to a reference PolyA (Truseq) approach on flash-frozen samples of the same tumors. We compared gene expression correlations and reproducibility. The Smarter Pico V3 ribodepletion approach appeared systematically the most comparable to Nanostring and Truseq (p < 0.001) and was a highly reproducible technique. In comparison with exome-capture and 3' kits, the Smarter appeared more comparable to Truseq (p < 0.001). Overall, our results suggest that the Smarter is the most robust RNA-Seq technique to study small FFPE samples and 3' Lexogen presents an interesting quality−price ratio for samples with less limiting quantities.

Performance of Next-Generation Sequencing for the Detection of Microsatellite Instability in Colorectal Cancer With Deficient DNA Mismatch Repair.

BACKGROUND & AIMS: Next-generation sequencing (NGS) was recently approved by the United States Food and Drug Administration to detect microsatellite instability (MSI) arising from defective mismatch repair (dMMR) in patients with metastatic colorectal cancer (mCRC) before treatment with immune checkpoint inhibitors (ICI). In this study, we aimed to evaluate and improve the performance of NGS to identify MSI in CRC, especially dMMR mCRC treated with ICI. METHODS: CRC samples used in this post hoc study were reassessed centrally for MSI and dMMR status using the reference methods of pentaplex polymerase chain reaction and immunohistochemistry. Whole-exome sequencing (WES) was used to evaluate MSISensor, the Food and Drug Administration-approved and NGS-based method for assessment of MSI. This was performed in (1) a prospective, multicenter cohort of 102 patients with mCRC (C1; 25 dMMR/MSI, 24 treated with ICI) from clinical trials NCT02840604 and NCT033501260, (2) an independent retrospective, multicenter cohort of 113 patients (C2; 25 mCRC, 88 non-mCRC, all dMMR/MSI untreated with ICI), and (3) a publicly available series of 118 patients with CRC from The Cancer Genome Atlas (C3; 51 dMMR/MSI). A new NGS-based algorithm, namely MSICare, was developed. Its performance for assessment of MSI was compared with MSISensor in C1, C2, and C3 at the exome level or after downsampling sequencing data to the MSK-IMPACT gene panel. MSICare was validated in an additional retrospective, multicenter cohort (C4) of 152 patients with new CRC (137 dMMR/MSI) enriched in tumors deficient in MSH6 (n = 35) and PMS2 (n = 9) after targeted sequencing of samples with an optimized set of microsatellite markers (MSIDIAG). RESULTS: At the exome level, MSISensor was highly specific but failed to diagnose MSI in 16% of MSI/dMMR mCRC from C1 (4 of 25; sensitivity, 84%; 95% confidence interval [CI], 63.9%-95.5%), 32% of mCRC (8 of 25; sensitivity, 68%; 95% CI, 46.5%-85.1%), and 9.1% of non-mCRC from C2 (8 of 88; sensitivity, 90.9%; 95% CI, 82.9%-96%), and 9.8% of CRC from C3 (5 of 51; sensitivity, 90.2%; 95% CI, 78.6%-96.7%). Misdiagnosis included 4 mCRCs treated with ICI, of which 3 showed an overall response rate without progression at this date. At the exome level, reevaluation of the MSI genomic signal using MSICare detected 100% of cases with true MSI status among C1 and C2. Further validation of MSICare was obtained in CRC tumors from C3, with 96.1% concordance for MSI status. Whereas misdiagnosis with MSISensor even increased when analyzing downsampled WES data from C1 and C2 with microsatellite markers restricted to the MSK-IMPACT gene panel (sensitivity, 72.5%; 95% CI, 64.2%-79.7%), particularly in the MSH6-deficient setting, MSICare sensitivity and specificity remained optimal (100%). Similar results were obtained with MSICare after targeted NGS of tumors from C4 with the optimized microsatellite panel MSIDIAG (sensitivity, 99.3%; 95% CI, 96%-100%; specificity, 100%). CONCLUSIONS: In contrast to MSISensor, the new MSICare test we propose performs at least as efficiently as the reference method, MSI polymerase chain reaction, to detect MSI in CRC regardless of the defective MMR protein under both WES and targeted NGS conditions. We suggest MSICare may rapidly become a reference method for NGS-based testing of MSI in CRC, especially in mCRC, where accurate MSI status is required before the prescription of ICI.

Immunophenotypic and molecular characterization of pancreatic neuroendocrine tumors producing serotonin.

Serotonin producing pancreatic neuroendocrine tumors (SP-PanNET) account for 0.58-1.4% of all pancreatic neuroendocrine tumors (PanNET). They may present with atypical symptoms, such as acute pancreatitis and are often radiologically characterized by main pancreatic duct dilatation. SP-PanNET are well differentiated neuroendocrine tumors (NET) distinct from classical PanNET by atypical serotonin secretion and abundant dense stroma deposition, like serotonin producing ileal NET leading in some cases to difficulties to reliably distinguish SP-PanNET from ileal NET metastases. The biology and molecular profile of SP-PanNET remain poorly characterized and the cell of origin within the pancreas is unclear. To address these questions, we analyzed a large cohort of SP-PanNET by immunohistochemistry (n = 29; ATRX, DAXX, MENIN, Islet1, PAX6, PDX1, ARX, CDX2), whole genome copy number array (Oncoscan™) and a large NGS panel (NovoPM™) (n = 10), FISH (n = 13) and RNA sequencing (n = 24) together with 21 ileal NET and 29 nonfunctioning PanNET (NF-PanNET). These analyses revealed a unique genomic profile with frequent isolated loss of chromosome 1 (14 cases-61%) and few pathogenic mutations (KMT2C in 2 cases, ARID1A in 1 case). Unsupervised RNAseq-based clustering showed that SP-PanNET were closer to NF-PanNET than ileal NET with an exclusive beta cell-like signature. SP-PanNET showed TGF-ß pathway activation signatures associated with extracellular matrix remodeling and similar signature were reproduced in vitro when pancreatic stellate cells were exposed to serotonin. SP-PanNET immunohistochemical profile resemble that of ileal NET except for PDX1 and PAX6 expression to a lesser extend suggesting that these two markers may be useful to diagnose SP-PanNET. Taken together, this suggests that SP-PanNET are a very specific PanNET entity with a peculiar biology leading to the characteristic fibrotic aspect.

DECONbench: a benchmarking platform dedicated to deconvolution methods for tumor heterogeneity quantification.

BACKGROUND: Quantification of tumor heterogeneity is essential to better understand cancer progression and to adapt therapeutic treatments to patient specificities. Bioinformatic tools to assess the different cell populations from single-omic datasets as bulk transcriptome or methylome samples have been recently developed, including reference-based and reference-free methods. Improved methods using multi-omic datasets are yet to be developed in the future and the community would need systematic tools to perform a comparative evaluation of these algorithms on controlled data. RESULTS: We present DECONbench, a standardized unbiased benchmarking resource, applied to the evaluation of computational methods quantifying cell-type heterogeneity in cancer. DECONbench includes gold standard simulated benchmark datasets, consisting of transcriptome and methylome profiles mimicking pancreatic adenocarcinoma molecular heterogeneity, and a set of baseline deconvolution methods (reference-free algorithms inferring cell-type proportions). DECONbench performs a systematic performance evaluation of each new methodological contribution and provides the possibility to publicly share source code and scoring. CONCLUSION: DECONbench allows continuous submission of new methods in a user-friendly fashion, each novel contribution being automatically compared to the reference baseline methods, which enables crowdsourced benchmarking. DECONbench is designed to serve as a reference platform for the benchmarking of deconvolution methods in the evaluation of cancer heterogeneity. We believe it will contribute to leverage the benchmarking practices in the biomedical and life science communities. DECONbench is hosted on the open source Codalab competition platform. It is freely available at: https://competitions.codalab.org/competitions/27453 .

Basal-like and classical cells coexist in pancreatic cancer revealed by single-cell analysis on biopsy-derived pancreatic cancer organoids from the classical subtype.

Pancreatic ductal adenocarcinoma (PDAC) is composed of stromal, immune, and cancerous epithelial cells. Transcriptomic analysis of the epithelial compartment allows classification into different phenotypic subtypes as classical and basal-like. However, little is known about the intra-tumor heterogeneity particularly in the epithelial compartment. Growing evidences suggest that this phenotypic segregation is not so precise and different cancerous cell types may coexist in a single tumor. To test this hypothesis, we performed single-cell transcriptomic analyses using combinational barcoding exclusively on epithelial cells from six different classical PDAC patients obtained by Endoscopic Ultrasound (EUS) with Fine Needle Aspiration (FNA). To purify the epithelial compartment, PDAC were grown as biopsy-derived pancreatic cancer organoids. Single-cell transcriptomic analysis allowed the identification of four main cell clusters present in different proportions in all tumors. Remarkably, although all these tumors were classified as classical, one cluster present in all corresponded to a basal-like phenotype. These results reveal an unanticipated high heterogeneity of pancreatic cancers and demonstrate that basal-like cells, which have a highly aggressive phenotype, are more widespread than expected.

Specific Genomic Alterations in High-Grade Pulmonary Neuroendocrine Tumours with Carcinoid Morphology.

INTRODUCTION: High-grade lung neuroendocrine tumours with carcinoid morphology have been recently reported; they may represent the thoracic counterparts of grade 3 digestive neuroendocrine tumours. We aimed to study their genetic landscape including analysis of tumoral heterogeneity. METHODS: Eleven patients with high-grade (>20% Ki-67 and/or >10 mitoses) lung neuroendocrine tumours with a carcinoid morphology were included. We analysed copy number variations, somatic mutations, and protein expression in 16 tumour samples (2 samples were available for 5 patients allowing us to study spatial and temporal heterogeneity). RESULTS: Genomic patterns were heterogeneous ranging from "quiet" to tetraploid, heavily rearranged genomes. Oncogene mutations were rare and most genetic alterations targeted tumour suppressor genes. Chromosomes 11 (7/11), 3 (6/11), 13 (4/11), and 6-17 (3/11) were the most frequently lost. Altered tumour suppressor genes were common to both carcinoids and neuroendocrine carcinomas, involving different pathways including chromatin remodelling (KMT2A, ARID1A, SETD2, SMARCA2, BAP1, PBRM1, KAT6A), DNA repair (MEN1, POLQ, ATR, MLH1, ATM), cell cycle (RB1, TP53, CDKN2A), cell adhesion (LATS2, CTNNB1, GSK3B) and metabolism (VHL). Comparative spatial/temporal analyses confirmed that these tumours emerged from clones of lower aggressivity but revealed that they were genetically heterogeneous accumulating "neuroendocrine carcinoma-like" genetic alterations through progression such as TP53/RB1 alterations. CONCLUSION: These data confirm the importance of chromatin remodelling genes in pulmonary carcinoids and highlight the potential role of TP53 and RB1 to drive the transformation in more aggressive high-grade tumours.