Developing a predictive model for metastatic potential in pancreatic neuroendocrine tumor.

CONTEXT: Pancreatic neuroendocrine tumors (PNETs) exhibit a wide range of behavior from localized disease to aggressive metastasis. A comprehensive transcriptomic profile capable of differentiating between these phenotypes remains elusive. OBJECTIVE: Use machine learning to develop predictive models of PNET metastatic potential dependent upon transcriptomic signature. METHODS: RNA-sequencing data were analyzed from 95 surgically-resected primary PNETs in an international cohort. Two cohorts were generated with equally balanced metastatic PNET composition. Machine learning was used to create predictive models distinguishing between localized and metastatic tumors. Models were validated on an independent cohort of 29 formalin-fixed, paraffin-embedded samples using NanoString nCounter®, a clinically-available mRNA quantification platform. RESULTS: Gene expression analysis identified concordant differentially expressed genes between the two cohorts. Gene set enrichment analysis identified additional genes that contributed to enriched biologic pathways in metastatic PNETs. Expression values for these genes were combined with an additional 7 genes known to contribute to PNET oncogenesis and prognosis, including ARX and PDX1. Eight specific genes (AURKA, CDCA8, CPB2, MYT1L, NDC80, PAPPA2, SFMBT1, ZPLD1) were identified as sufficient to classify the metastatic status with high sensitivity (87.5% - 93.8%) and specificity (78.1% - 96.9%). These models remained predictive of the metastatic phenotype using NanoString nCounter® on the independent validation cohort, achieving a median AUROC of 0.886. CONCLUSIONS: We identified and validated an eight-gene panel predictive of the metastatic phenotype in PNETs, which can be detected using the clinically-available NanoString nCounter® system. This panel should be studied prospectively to determine its utility in guiding operative versus non-operative management.

Medullary Thyroid Cancer: Single Institute Experience over Three Decades and Risk Factors for Recurrence.

CONTEXT: Medullary thyroid cancer has a historic recurrence rate up to 50%, and surgery remains the only cure. OBJECTIVE: This study aims to assess factors related to recurrence and metastatic spread in MTC. DESIGN: Retrospective chart review was performed from 1990-2023. Descriptive analysis and regression models were used for analysis. SETTING: Single specialized tertiary care referral center. PATIENTS: 68 patients with MTC, who underwent surgery, were included. MAIN OUTCOME MEASURE: Recurrence. RESULTS: Mean age at diagnosis was 54.9years(42.2-64.1), 65%(n=44) females. Lymph node and distant metastases were found in 24%(n=16) and 4%(n=3), respectively. RET mutations were present in 52%(n=35): MTC risk levels Highest 6%, High 7%, and Moderate 39%. Mean tumor size was 1.9cm(1.2-3.2) and mean preoperative calcitonin was 504.4pg/mL(133.2-1833.8). Total thyroidectomy(TT) was performed in 10 patients, TT+central neck dissection(CND) in 28, and TT+CND+lateral neck dissection(LND) in 25. On final pathology, 40% had positive central nodes and 25% had positive lateral nodes. Recurrence was 22%, median follow-up 4.7years(1.2-28.0). Male gender(HR=5.81, p=0.021), positive lateral neck nodes(HR 8.10, p=0.011) and high/highest MTC risk level RET mutations(HR 8.66, p=0.004) were significantly associated with recurrence. Preoperative calcitonin>2,175 pg/mL was a strong predictor for distant metastasis(AUC0.893) and a good predictor for lateral neck disease(AUC0.706). Extent of surgery was not significantly associated with recurrence(p=0.634). CONCLUSION: One of 4 patients undergoing surgery for MTC will recur. Risk factors associated with recurrence are male gender, lateral LN metastasis and high/highest MTC risk level mutations, but not necessarily surgery type. Preoperative calcitonin>2,175 pg/mL is suggestive of advanced disease and should prompt further evaluation.

Clinical Significance of Pheochromocytoma Size on the Timing and Extent of Surgery.

INTRODUCTION: Elevated metanephrine and catecholamine levels 3-fold upper limit of normal (ULN) are diagnostic for pheochromocytoma. We sought to determine whether size correlates with biochemical activity or symptoms which could guide timing of surgery. METHODS: Data from consecutive patients undergoing adrenalectomy for pheochromocytoma at our institution over a 10-year period were retrospectively collected. These included maximal lesion diameter on preoperative imaging, plasma/urine metanephrine and/or catecholamine levels, demographic variables and presence of typical paroxysmal symptoms. Receiver operating characteristic curves were used to assess predictive accuracy. RESULTS: Sixty-three patients were included in the analysis (41 females and 22 males). Median age was 56 (43, 69) years. Due to various referring practices, 31 patients had documented 24-h urine metanephrine, 26 had 24-h urine catecholamine, and 52 had fractionated plasma metanephrine levels available for review. Values were converted to fold change compared to ULN and the maximum of all measured values was used for logistic regression. Median tumor size was 3.40 (2.25, 4.55) cm in greatest dimension. Tumor size at which pheochromocytoma produced > 3-fold ULN was ≥2.3 cm (AUC of 0.84). Biochemical activity increased with doubling tumor size (odds ratio = 8, P = 0.0004) or ≥ 1 cm increase in tumor size (odds ratio = 3.03, P = 0.001). 40 patients had paroxysmal symptoms, but there was no significant correlation between tumor size/biochemical activity and symptoms. CONCLUSIONS: In our study, tumor size directly correlated with the degree of biochemical activity and pheochromocytomas ≥2.3 cm produced levels 3 times ULN. These findings may allow clinicians to adjust timing of operative intervention.

Single center outcomes from parenchymal-sparing resections and microwave ablations for neuroendocrine tumor liver metastases.

BACKGROUND: Reported outcomes after surgical debulking in patients with advanced neuroendocrine tumor liver metastases (NETLM) are sparse. METHODS: NETLM patients that underwent surgical debulking from 2019 to 2021 were reviewed. Trends in perioperative liver function, complications, symptom response, and progression-free survival were examined. RESULTS: 1069 liver lesions were debulked from 53 patients using a combination of parenchymal-sparing resections (PSR) and ultrasound-guided microwave ablations (MWA). Post-operative transaminitis and thrombocytopenia were common, and severity correlated with increasing number of lesions. Laboratory markers for synthetic liver function did not differ according to the number of lesions debulked. 13% of patients sustained a Clavien-Dindo grade 3 or 4 complication which was not associated with the number of lesions targeted. All patients with preoperative symptoms had improvement after surgery. Median time to progression was 10.9 months. CONCLUSIONS: PSR with MWA for large numbers of NETLM is safe and effective for symptom control and does not affect synthetic liver function.

Clinical utility of a microRNA classifier in cytologically indeterminate thyroid nodules with RAS mutations: A multi-institutional study.

BACKGROUND: Molecular testing guides the management of cytologically indeterminate thyroid nodules. We evaluated the real-world clinical benefit of a commercially available thyroid mutation panel plus microRNA risk classifier in classifying RAS-mutated nodules. METHODS: We performed a subgroup analysis of the results of molecular testing of Bethesda III/IV nodules using the ThyGenX/ThyGeNEXT-ThyraMIR platform at 3 tertiary-care centers between 2017 and 2021, defining a positive result as 10% or greater risk of malignancy. RESULTS: We identified 387 nodules from 375 patients (70.7% female, median age 59.3 years) who underwent testing. Positive nodules (32.3%) were associated with increased surgical intervention (74.4% vs 14.9%, P < .0001) and carcinoma on surgical pathology (46.4% vs 3.4%, P < .0001) compared to negative modules. RAS mutations were the most common mutations, identified in 71 of 380 (18.7%) nodules, and were classified as ThyraMIR- (28 of 71; 39.4%) or ThyraMIR+ (43 of 71; 60.6%). Among RAS-mutated nodules, there was no significant difference in operative rate (P = .2212) or carcinoma diagnosis (P = .6277) between the ThyraMIR+ and ThyraMIR- groups, and the sensitivity, specificity, negative predictive value, and positive predictive value of ThyraMIR were 64.7%, 34.8%, 40.0%, and 59.5%, respectively. CONCLUSION: Although testing positive is associated with malignancy in surgical pathology, the ThyraMIR classifier failed to differentiate between benign and malignant RAS-mutated nodules. Diagnostic lobectomy should be considered for RAS-mutated nodules, regardless of microRNA expression status.

Number of Hepatic Resections and Association with Complication after Metastatic Neuroendocrine Tumor Debulking: An American College of Surgeons NSQIP Database Study.

BACKGROUND: Neuroendocrine tumor liver metastases (NELM) are a major source of morbidity and mortality in neuroendocrine tumor patients and can be treated with hepatic debulking surgery (HDS). This study aims to identify variables associated with postoperative morbidity in NELM HDS. STUDY DESIGN: This analysis uses the American College of Surgeons NSQIP targeted hepatectomy-specific Participant User File from 2014 to 2020. Surgeries were grouped by number of hepatic resections performed (1 to 5, 6 to 10, greater than 10). Multivariable logistic regression and matching were used to identify prognostic factors of morbidity. RESULTS: A total of 1,163 patients were included. A total of 1,011 (87%) had 1 to 5 hepatic resections, 101 (8.7%) had 6 to 10 resections, and 51 (4.4%) had greater than 10 resections. The overall complication rate was 35%, and surgical and medical complications reached 30% and 13%, respectively. Mortality occurred in 11 patients (0.9%). Significantly higher rates of any (34% vs 35% vs 53%, p = 0.021) and surgical complications (29% vs 28% vs 49%, p = 0.007) were noted for those undergoing greater than 10 resections (1 to 5 resections, 6 to 10, greater than 10). "Bleeding requiring transfusion" (p < 0.0001) occurred more frequently in the greater than 10 resection group. On multivariable logistic regression, greater than 10 resections were an independent risk factor for any (odds ratio [OR] 2.53, p = 0.002; OR 2.52, p =. 0013) and surgical (OR 2.53, p = 0.003; OR 2.88, p = 0.005) complications compared with 1 to 5 resections and 6 to 10 resections, respectively. Medical complications (OR 2.34, p = 0.020) and length of stay greater than 5 days (OR 1.98, p = 0.032) were also increased with greater than 10 vs 1 to 5 resections. CONCLUSIONS: As reported by NSQIP, NELM HDS were performed safely with low mortality. However, more hepatic resections, especially greater than 10, were associated with increased postoperative morbidity and length of stay.

Surgery for metastatic pancreatic neuroendocrine tumors: a narrative review.

Background and Objective: Pancreatic neuroendocrine tumors (PanNETs) are derived from the islet cells of the pancreas and have been increasing in incidence. Most of these tumors are nonfunctional although some can secrete hormones and lead to hormone-specific clinical syndromes. Surgery is the mainstay of treatment for localized tumors, however, surgical resection is controversial in metastatic PanNETs. This narrative review seeks to summarize the current literature surrounding surgery, specifically in the controversial area of metastatic PanNETs, review current treatment paradigms, and understand the benefits of surgery in this group of patients. Methods: Authors searched PubMed using the terms "surgery pancreatic neuroendocrine tumor", "metastatic neuroendocrine tumor", and "liver debulking neuroendocrine tumor" from January 1990 to June 2022. Only English language publications were considered. Key Content and Findings: There is no consensus among the leading specialty organizations regarding surgery for metastatic PanNETs. When considering surgery for metastatic PanNETs, tumor grade and morphology, location of the primary tumor, extra-hepatic or extra-abdominal disease, as well as liver tumor burden and metastatic distribution should be considered. Because the liver is the most common site of metastasis and liver failure is the most common cause of death in patients with hepatic metastases, attention is centered here on debulking and other ablative techniques. Liver transplantation is rarely used for hepatic metastases but could be beneficial in a small subset of patients. Retrospective studies have demonstrated improvement in survival and symptoms after surgery for metastatic disease, but the lack of prospective randomized control trials significantly limits analysis of surgical benefits in patients with metastatic PanNETs. Conclusions: Surgery is the standard of care for localized PanNETs, while it remains controversial in metastatic disease. Many studies have shown a survival and symptomatic benefit to surgery and liver debulking in select groups of patients. However, most of the studies on which recommendations are based in this population are retrospective in nature and are subject to selection bias. This presents an opportunity for future investigation.

Light It Up! The Use of DOTATATE in Diagnosis and Treatment of Neuroendocrine Neoplasms.

Radiolabeled somatostatin analogs are increasingly used in the diagnosis and treatment of neuroendocrine tumors. Diagnostic imaging with 68Ga-DOTATATE PET/CT has demonstrated the improved sensitivity in detecting primary and metastatic neuroendocrine lesions compared with conventional imaging and prior generation somatostatin receptor imaging. Peptide receptor radionuclide therapy with 177Lu-DOTATATE is now frequently included in the management of neuroendocrine neoplasms, with prospective randomized control studies demonstrating its beneficial impact on survival and quality of life. Nonetheless, peptide rector radionuclide therapy is still considered palliative rather than curative and may be accompanied by adverse effects.

A machine-learning algorithm for distinguishing malignant from benign indeterminate thyroid nodules using ultrasound radiomic features.

Background: Ultrasound (US)-guided fine needle aspiration (FNA) cytology is the gold standard for the evaluation of thyroid nodules. However, up to 30% of FNA results are indeterminate, requiring further testing. In this study, we present a machine-learning analysis of indeterminate thyroid nodules on ultrasound with the aim to improve cancer diagnosis. Methods: Ultrasound images were collected from two institutions and labeled according to their FNA (F) and surgical pathology (S) diagnoses [malignant (M), benign (B), and indeterminate (I)]. Subgroup breakdown (FS) included: 90 BB, 83 IB, 70 MM, and 59 IM thyroid nodules. Margins of thyroid nodules were manually annotated, and computerized radiomic texture analysis was conducted within tumor contours. Initial investigation was conducted using five-fold cross-validation paradigm with a two-class Bayesian artificial neural networks classifier, including stepwise feature selection. Testing was conducted on an independent set and compared with a commercial molecular testing platform. Performance was evaluated using receiver operating characteristic analysis in the task of distinguishing between malignant and benign nodules. Results: About 1052 ultrasound images from 302 thyroid nodules were used for radiomic feature extraction and analysis. On the training/validation set comprising 263 nodules, five-fold cross-validation yielded area under curves (AUCs) of 0.75 [Standard Error (SE) = 0.04; P < 0.001 ] and 0.67 (SE = 0.05; P = 0.0012 ) for the classification tasks of MM versus BB, and IM versus IB, respectively. On an independent test set of 19 IM/IB cases, the algorithm for distinguishing indeterminate nodules yielded an AUC value of 0.88 (SE = 0.09; P < 0.001 ), which was higher than the AUC of a commercially available molecular testing platform (AUC = 0.81, SE = 0.11; P < 0.005 ). Conclusion: Machine learning of computer-extracted texture features on gray-scale ultrasound images showed promising results classifying indeterminate thyroid nodules according to their surgical pathology.

Loss of MEN1 function impairs DNA repair capability of pancreatic neuroendocrine tumors.

Somatic MEN1 mutations occur in up to 50% of pancreatic neuroendocrine tumors (PanNETs). Clinical studies have shown that radiation therapy (IR) is effective in a subset of PanNETs, but it remains unclear why some patients respond better to IR than others. Herein, we study whether MEN1 loss of function increases radiosensitivity of PanNETs and determine its effect on DNA double-strand break (DSB) repair. After creating a MEN1 knockout PanNET cell line, we confirmed reduced DSB repair capacity in MEN1-deficient cells and linked these findings to a defect in homologous recombination, as well as reduced BRCA2 expression levels. Consistent with this model, we found that MEN1 mutant cells displayed increased sensitivity to the highly trapping poly (ADP-ribose) polymerase (PARP) 1 inhibitor talazoparib in vitro. Our results suggest that combining IR with PARP inhibition may be beneficial in patients with PanNETs and MEN1 loss of function.

Epigenetic Dysregulation of 5-hydroxymethylcytosine in Well-Differentiated Pancreatic Neuroendocrine Tumors.

Dysregulation of epigenetic mechanisms, reflected by loss of expression of 5-hydroxymethylcytosine (5-hmC) is being increasingly recognized as a marker of aggressive behavior in several neoplasms; however, the role of such epigenetic modifiers in pancreatic neuroendocrine tumors (PanNETs) has not been studied. Annotated cohort of 60 PanNETs was evaluated for 5-hmC expression using immunohistochemistry. Univariable and multivariable analyses were performed. To determine intratumor heterogeneity of 5-hmC expression, 26 additional synchronous metastatic deposits of PanNETs from 8 patients were evaluated for 5-hmC expression. 5-hmC level showed significant association with the presence of distant metastases (P=0.02), female sex (P=0.04), and Ki-67 proliferation index (P=0.002). A multivariate model created using the stepwise logistic regression analysis showed the presence of nodal metastases (odds ratio=6.15), lymphovascular invasion (odds ratio=4.07) and lack of 5-hmC expression (odds ratio=5.34) were predictive of the risk of distant metastasis in PanNETs with a c-statistic of 0.845. Epigenetic intratumoral heterogeneity of 5-hmC expression was seen in 37.5% cases (3/8). Our work provides evidence that epigenetic regulators are involved in the pathobiology of PanNETs and immunohistochemical analysis of 5-hmC may be able to refine prognostic evaluation of these tumors.

Management recommendations for pancreatic manifestations of von Hippel-Lindau disease.

Von Hippel-Lindau disease (VHL) is a multineoplasm inherited disease manifesting with hemangioblastoma of the central nervous system and retina, adrenal pheochromocytoma, renal cell carcinoma, pancreatic neuroendocrine tumors and cysts, and neoplasms/cysts of the ear, broad ligament, and testicles. During 2018-2020, the VHL Alliance gathered several committees of experts in the various clinical manifestations of VHL to review the literature, gather the available evidence on VHL, and develop recommendations for patient management. The current report details the results of the discussion of a group of experts in the pancreatic manifestations of VHL along with their proposed recommendations for the clinical surveillance and management of patients with VHL. The recommendations subcommittee performed a comprehensive systematic review of the literature and conducted panel discussions to reach the current recommendations. The level of evidence was defined according to the Shekelle variation of the Grading of Recommendations, Assessment, Development, and Evaluation grading system. The National Comprehensive Cancer Network Categories of Evidence and Consensus defined the committee members' interpretation of the evidence and degree of consensus. The recommendations encompass the main aspects of VHL-related pancreatic manifestations and their clinical management. They are presented in a clinical orientation, including general planning of screening and surveillance for pancreatic neuroendocrine tumors, utility of biochemical biomarkers, the optimal choice for imaging modality, indirect risk stratification, indications for tissue sampling of VHL-related pancreatic neuroendocrine tumors, and interventions. These recommendations are designed to serve as the reference for all aspects of the screening, surveillance, and management of VHL-related pancreatic manifestations.

Sunitinib-Loaded Chondroitin Sulfate Hydrogels as a Novel Drug-Delivery Mechanism for the Treatment of Pancreatic Neuroendocrine Tumors.

BACKGROUND: Pancreatic neuroendocrine tumors (PanNETs) are increasingly common. Experts debate whether small tumors should be resected. Tumor destruction via injection of cytotoxic agents could offer a minimal invasive approach to this controversy. We hypothesize that a new drug delivery system comprising chondroitin sulfate (CS) hydrogels loaded with sunitinib (SUN) suppresses tumor growth in PanNET cells. METHODS: Injectable hydrogels composed of CS modified with methacrylate groups (MA) were fabricated and loaded with SUN. Loading target was either 200 µg (SUN200-G) or 500 µg (SUN500-G) as well as sham hydrogel with no drug loading (SUN0-G). SUN release from hydrogels was monitored in vitro over time and cytotoxicity induced by the released SUN was evaluated using QGP-1 and BON1 PanNET cell lines. QGP-1 xenografts were developed in 35 mice and directly injected with 25 µL of either SUN200-G, SUN500-G, SUN0-G, 100 µL of Sunitinib Malate (SUN-inj), or given 40 mg/kg/day oral sunitinib (SUN-oral). RESULTS: SUN-loaded CSMA hydrogel retained complete in vitro cytotoxicity toward the QGP-1 PanNET and BON-1 PanNET cell lines for 21 days. Mouse xenograft models with QGP-1 PanNETs showed a significant delay in tumor growth in the SUN200/500-G, SUN-inj and SUN-oral groups compared with SUN0-G (p = 0.0014). SUN500-G hydrogels induced significantly more tumor necrosis than SUN0-G (p = 0.04). There was no difference in tumor growth delay between SUN200/500G, SUN-inj, and SUN-oral. CONCLUSIONS: This study demonstrates that CSMA hydrogels loaded with SUN suppress PanNETs growth. This drug delivery could approach represents a novel way to treat PanNETs and other neoplasms via intratumoral injection.

Preoperative serum chromogranin-a is predictive of survival in locoregional jejuno-ileal small bowel neuroendocrine tumors.

BACKGROUND: Small bowel neuroendocrine tumors (SB-NET) frequently metastasize to regional lymphatic or distant sites. Although most prognostication of SB-NET focuses on lymph node involvement, findings from studies of neuroendocrine tumors from other primary sites have suggested that preoperative serum chromogranin-A (CgA) levels may provide a more accurate metric. METHODS: Using the National Cancer Database (2004-2016), we analyzed patients with locoregional SB-NET who underwent curative resection including an adequate lymphadenectomy (n = 1,274). A statistically optimized cut-point was used to dichotomize CgA cohort based on preoperative serum CgA levels. RESULTS: We determined that a CgA ≥139 ng/mL identified patients with significantly shorter estimated mean overall survival (6.6 years vs 7.6 years, log-rank P = .00001). These patients were also older (63 vs 57 years, P < .001) and had higher rates of poorly differentiated tumors (2.1% vs 0.7%, P = .04) or primary tumors >1 cm (88.2% vs 79.2%, P = .001). Clinical features associated with shorter overall survival included preoperative CgA ≥139 ng/mL (HR = 2.19, 95% CI 1.22-3.92; P = .009), age at diagnosis (HR = 1.06, 95% CI 1.03-1.09; P < .001), Charlson-Deyo score ≥2 (HR = 3.93, 95% CI 1.71-9.01; P = .001), and poorly differentiated tumors (HR = 11.22, 95% CI 4.16-30.24; P < .001). Neither lymph node metastasis nor T-stage were independently associated with shorter overall survival in patients with locoregional SB-NET. CONCLUSION: Elevated preoperative serum CgA is an adverse prognostic marker associated with shorter overall survival in patients with locoregional SB-NET.

Total Thyroidectomy vs Thyroid Lobectomy for Localized Papillary Thyroid Cancer in Children: A Propensity-Matched Survival Analysis.

BACKGROUND: Current guidelines recommend total thyroidectomy (TT) and radioablation for most papillary thyroid cancer (PTC) in children. These guidelines have been criticized as aggressive, especially for early-stage PTC, as it likely does not influence patient survival and results in life-long thyroid hormone replacement. We sought to study whether the extent of thyroidectomy (TT vs thyroid lobectomy [TL]) influences overall and disease-specific survival in children with localized PTC. METHODS: The National Cancer Database and the Surveillance, Epidemiology, and End Results registries were queried. Patients 18 years or younger with low-risk PTC between 2004 and 2016 were included. Using a 1:1 propensity score matching, patients who underwent TT were matched for age, sex, race, year of diagnosis, and tumor size with a similar cohort of patients who underwent TL. Primary end points were overall survival and disease-specific survival. RESULTS: There were 3,500 patients identified as surgically treated for PTC, of which 1,325 patients met inclusion criteria for matching. Three hundred and twenty-six patients were matched. One hundred and sixty-three patients had TT; 140 were female and mean age was 16 years (interquartile range [IQR] 13 to 17 years). One hundred and sixty-three patients had TL; 140 were female and mean age was 16 years (IQR 14 to 17 years). Median follow-up was 5.0 years (IQR 2.8 to 8 years) and 8.3 years (IQR 3.6 to 14.4 years) in the National Cancer Database and Surveillance, Epidemiology, and End Results cohorts, respectively. There was no statistically significant difference in overall survival or disease-specific survival in patients with PTC < 4 cm, regardless of whether patients underwent TT or TL (p = 0.32 for National Cancer Database registry and p = 0.67 for Surveillance, Epidemiology, and End Results registry). CONCLUSIONS: This study suggests that the extent of thyroidectomy does not influence survival for pediatric patients with early-stage PTC and that TL might be adequate in this patient population.

Deep learning prediction of BRAF-RAS gene expression signature identifies noninvasive follicular thyroid neoplasms with papillary-like nuclear features.

Noninvasive follicular thyroid neoplasms with papillary-like nuclear features (NIFTP) are follicular-patterned thyroid neoplasms defined by nuclear atypia and indolent behavior. They harbor RAS mutations, rather than BRAFV600E mutations as is observed in papillary thyroid carcinomas with extensive follicular growth. Reliably identifying NIFTPs aids in safe therapy de-escalation, but has proven to be challenging due to interobserver variability and morphologic heterogeneity. The genomic scoring system BRS (BRAF-RAS score) was developed to quantify the extent to which a tumor's expression profile resembles a BRAFV600E or RAS-mutant neoplasm. We proposed that deep learning prediction of BRS could differentiate NIFTP from other follicular-patterned neoplasms. A deep learning model was trained on slides from a dataset of 115 thyroid neoplasms to predict tumor subtype (NIFTP, PTC-EFG, or classic PTC), and was used to generate predictions for 497 thyroid neoplasms within The Cancer Genome Atlas (TCGA). Within follicular-patterned neoplasms, tumors with positive BRS (RAS-like) were 8.5 times as likely to carry an NIFTP prediction than tumors with negative BRS (89.7% vs 10.5%, P < 0.0001). To test the hypothesis that BRS may serve as a surrogate for biological processes that determine tumor subtype, a separate model was trained on TCGA slides to predict BRS as a linear outcome. This model performed well in cross-validation on the training set (R2 = 0.67, dichotomized AUC = 0.94). In our internal cohort, NIFTPs were near universally predicted to have RAS-like BRS; as a sole discriminator of NIFTP status, predicted BRS performed with an AUC of 0.99 globally and 0.97 when restricted to follicular-patterned neoplasms. BRAFV600E-mutant PTC-EFG had BRAFV600E-like predicted BRS (mean -0.49), nonmutant PTC-EFG had more intermediate predicted BRS (mean -0.17), and NIFTP had RAS-like BRS (mean 0.35; P < 0.0001). In summary, histologic features associated with the BRAF-RAS gene expression spectrum are detectable by deep learning and can aid in distinguishing indolent NIFTP from PTCs.

Indoleamine 2,3-Dioxygenase-1 Expression in Adrenocortical Carcinoma.

BACKGROUND: Indoleamine 2,3-dioxygenase 1 (IDO-1) is overexpressed in many human carcinomas and a successful target for therapy in mouse models. Prognosis of patients with advanced adrenocortical carcinoma (ACC) is poor due to the lack of effective treatments, and new therapies are therefore needed. Herein, we investigate whether IDO-1 is expressed in human ACC tissues. METHODS: 53 tissue samples from patients with ACC, adrenal adenoma (AA), adrenocortical tumors (ACTs), and normal adrenal were identified. Immunohistochemistry was performed on formalin-fixed, paraffin-embedded slides for IDO-1. Samples were scored for cytoplasmic staining as per intensity and the percent of positive cells and for stromal staining by percent of positive cells. Tumor characteristics, PD-L1, PDL-2, and CD-8+ T-lymphocyte expression were also determined. RESULTS: Samples from 32 ACC, 3 ACT, 15 AA, and 3 normal adrenal were analyzed. IDO-1 was expressed in tumor tissue in 22 of 32 ACC samples, compared with 8 of 15 AA sample (P = 0.344). IDO-1 expression was significantly increased in stromal tissue of ACC samples (16 of 33), compared with AA samples (0 of 15) (P = 0.001). IDO-1 expression in ACC and AA samples was associated with PD-L2 expression (P = 0.034). IDO-1 expression in ACC stromal tissue was associated with CD8+ T-lymphocyte infiltration (P = 0.028). CONCLUSIONS: IDO-1 is expressed in a majority of ACC samples. Its expression in tumor tissue is associated with PD-L2 expression, and expression in stroma is associated with CD8+ cell infiltration. IDO-1 inhibition, alone or in combination with PD-1 inhibition, could therefore be an interesting target in treatment of ACC.

Risk factors associated with positive resection margins in patients with adrenocortical carcinoma.

BACKGROUND: Positive resection margins are associated with worse survival after surgery for adrenocortical carcinoma (ACC). We aimed to identify risk factors for positive margins post-resection. METHODS: The NCDB was queried for ACC patients from 2006 to 2015. Patients with positive versus negative resection margins post-surgery were compared using Chi-square tests. Survival based on adjuvant treatment was assessed using Kaplan-Meier curves. RESULTS: 1,973 patients with ACC were identified, 217 (11.0%) with positive margins. Multivariable analysis identified extra-adrenal extension (HR 4.92, p < 0.001), lymph node metastases (HR 2.64, p = 0.001), and distant metastases (HR 1.53, p = 0.03) as risk factors for positive margins. No significant difference in margin status existed between patients who had an open versus minimally invasive procedure (p = 0.6). Positive margin patients receiving adjuvant radiation (p = 0.007) or combined chemo-radiation (p = 0.001) had the longest survival. CONCLUSION: No modifiable risk factors were identified, but patients with positive margins receiving adjuvant radiation or chemo-radiation had the longest survival.