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.

Serotonin immunoreactive pancreatic neuroendocrine neoplasm associated with main pancreatic duct dilation: a recognizable entity with excellent long-term outcome.

OBJECTIVES: Dilatation of the main pancreatic duct (MPD) is rare in pancreatic neuroendocrine neoplasm (panNEN) and may be due to different mechanisms. We compared the imaging and pathological characteristics as well as the outcome after resection of positive (S+) and negative (S-) serotonin immunoreactive panNENs causing MPD dilatation. METHODS: This retrospective study included patients with panNEN, with MPD dilatation (≥ 4 mm) on preoperative CT/MRI and resected between 2005 and 2019. Clinical, radiological, and pathological features were compared between S+ and S- panNENs. Imaging features associated with S+ panNEN were identified using logistic regression analysis. The diagnostic performance of imaging for the differentiation of S+ and S- panNENs was assessed by ROC curve analysis. Recurrence-free survival (RFS) was compared between the two groups. RESULTS: The final population of 60 panNENs included 20/60 (33%) S+ panNENs. S+ panNENs were smaller (median 12.5 mm vs. 33 mm; p < 0.01), more frequently hyperattenuating/intense on portal venous phase at CT/MRI (95% vs. 25%, p < 0.01), and presented with more fibrotic stroma on pathology (60.7 ± 16% vs. 40.7 ± 12.8%; p < 0.01) than S- panNENs. Tumor size was the only imaging factor associated with S+ panNEN on multivariate analysis. A tumor size ≤ 20 mm had 95% sensitivity and 90% specificity for the diagnosis of S+ panNEN. Among 52 patients without synchronous liver metastases, recurrence occurred in 1/20 (5%) with S+ panNEN and 18/32 (56%) with S- panNEN (p < 0.01). Median RFS was not reached in S+ panNENs and was 31.3 months in S- panNENs (p < 0.01). CONCLUSIONS: In panNENs with MPD dilatation, serotonin positivity is associated with smaller size, extensive fibrotic stroma, and better long-term outcomes. KEY POINTS: • S+ panNENs showed a higher percentage of fibrotic stroma, higher microvessel density, and lower proliferation index (Ki-67) compared to S- panNENs. • Radiologically, S+ panNENs causing dilatation of the MPD were characterized by a small size (< = 20 mm) and a persistent enhancement on portal phase on both CT and MRI. • Patients with S+ panNENs presented with longer RFS when compared to those with S- panNENs.

Multicenter automatic detection of invasive carcinoma on breast whole slide images.

Breast cancer is one of the most prevalent cancers worldwide and pathologists are closely involved in establishing a diagnosis. Tools to assist in making a diagnosis are required to manage the increasing workload. In this context, artificial intelligence (AI) and deep-learning based tools may be used in daily pathology practice. However, it is challenging to develop fast and reliable algorithms that can be trusted by practitioners, whatever the medical center. We describe a patch-based algorithm that incorporates a convolutional neural network to detect and locate invasive carcinoma on breast whole-slide images. The network was trained on a dataset extracted from a reference acquisition center. We then performed a calibration step based on transfer learning to maintain the performance when translating on a new target acquisition center by using a limited amount of additional training data. Performance was evaluated using classical binary measures (accuracy, recall, precision) for both centers (referred to as "test reference dataset" and "test target dataset") and at two levels: patch and slide level. At patch level, accuracy, recall, and precision of the model on the reference and target test sets were 92.1% and 96.3%, 95% and 87.8%, and 73.9% and 70.6%, respectively. At slide level, accuracy, recall, and precision were 97.6% and 92.0%, 90.9% and 100%, and 100% and 70.8% for test sets 1 and 2, respectively. The high performance of the algorithm at both centers shows that the calibration process is efficient. This is performed using limited training data from the new target acquisition center and requires that the model is trained beforehand on a large database from a reference center. This methodology allows the implementation of AI diagnostic tools to help in routine pathology practice.

Multicenter Evaluation of the Idylla GeneFusion in Non-Small-Cell Lung Cancer.

Targeted therapy in lung cancer requires the assessment of multiple oncogenic driver alterations, including fusion genes. This retrospective study evaluated the Idylla GeneFusion prototype, an automated and ease-of-use (<2 minutes) test, with a short turnaround time (3 hours) to detect fusions involving ALK, ROS1, RET, and NTRK1/2/3 genes and MET exon 14 skipping. This multicenter study (18 centers) included 313 tissue samples from lung cancer patients with 97 ALK, 44 ROS1, 20 RET, and 5 NTRKs fusions, 32 MET exon 14 skipping, and 115 wild-type samples, previously identified with reference methods (RNA-based next-generation sequencing/fluorescence in situ hybridization/quantitative PCR). Valid results were obtained for 306 cases (98%), overall concordance between Idylla and the reference methods was 89% (273/306); overall sensitivity and specificity were 85% (165/193) and 96% (108/113), respectively. Discordances were observed in 28 samples, where Idylla did not detect the alteration identified by the reference methods; and 5 samples where Idylla identified an alteration not detected by the reference methods. All of the ALK-, ROS1-, and RET-specific fusions and MET exon 14 skipping identified by Idylla GeneFusion were confirmed by reference method. To conclude, Idylla GeneFusion is a clinically valuable test that does not require a specific infrastructure, allowing a rapid result. The absence of alteration or the detection of expression imbalance only requires additional testing by orthogonal methods.