Automated assessment of Ki-67 proliferation index in neuroendocrine tumors by deep learning.

The Ki-67 proliferation index (PI) is a prognostic factor in neuroendocrine tumors (NETs) and defines tumor grade. Analysis of Ki-67 PI requires calculation of Ki-67-positive and Ki-67-negative tumor cells, which is highly subjective. To overcome this, we developed a deep learning-based Ki-67 PI algorithm (KAI) that objectively calculates Ki-67 PI. Our study material consisted of NETs divided into training (n = 39), testing (n = 124), and validation (n = 60) series. All slides were digitized and processed in the Aiforia® Create (Aiforia Technologies, Helsinki, Finland) platform. The ICC between the pathologists and the KAI was 0.89. In 46% of the tumors, the Ki-67 PIs calculated by the pathologists and the KAI were the same. In 12% of the tumors, the Ki-67 PI calculated by the KAI was 1% lower and in 42% of the tumors on average 3% higher. The DL-based Ki-67 PI algorithm yields results similar to human observers. While the algorithm cannot replace the pathologist, it can assist in the laborious Ki-67 PI assessment of NETs. In the future, this approach could be useful in, for example, multi-center clinical trials where objective estimation of Ki-67 PI is crucial.

IDH1 Expression via the R132H Mutation-Specific Antibody in Adrenocortical Neoplasias-Prognostic Impact in Carcinomas.

CONTEXT: Mutations to isocitrate dehydrogenase (IDH) appear to play a prognostic or predictive role in several neoplasias. Immunohistochemical staining designed to detect a specific R132H mutation to IDH1 showed expression in the normal adrenal cortex, raising interest to study the potential role of IDH1 in the pathogenesis of adrenocortical tumors. OBJECTIVE: The objective of this work is to study the role of IDH1 and its mutations in adrenocortical tumors. DESIGN AND PATIENTS: IDH1 R132H immunohistological staining was performed on a cohort of 197 adrenocortical tumors. The exon of the IDH1 gene was sequenced in 16 tumors. RESULTS: Positive IDH1 R132H immunohistochemical staining correlated with a better prognosis among patients with a malignant adrenocortical tumor. However, IDH1 R132H immunohistochemistry did not distinguish between local and metastasized tumors. We were unable to identify IDH1 mutations among our adrenocortical tumors using a targeted next-generation sequencing panel or via exon sequencing. CONCLUSIONS: Among adrenocortical carcinomas, IDH1 R132H immunopositivity correlated with a better prognosis. Thus, IDH1 R132H immunohistochemical staining could serve as a prognostic or as a potential predictive marker in adrenocortical carcinomas. Further research is needed to identify the possible alterations in IDH1 that could explain our findings, because we identified no known mutations to the IDH1 gene.

Adrenocortical carcinoma: presentation and outcome of a contemporary patient series.

BACKGROUND: Adrenocortical carcinoma (ACC) is a rare endocrine carcinoma with poor 5-year survival rates of < 40%. According to the literature, ACC is rarely an incidental imaging finding. However, presentation, treatment and outcome may differ in modern series. DESIGN AND METHODS: We studied all patients (n = 47, four children) from a single centre during years 2002-2018. We re-evaluated radiologic and histopathological findings and assessed treatments and outcome. We searched for possible TP53 gene defects and assessed nationwide incidence of ACC. RESULTS: In adults, incidental radiologic finding led to diagnosis in 79% at median age of 61 years. ENSAT stage I, II, III and IV was 19%, 40%, 19% and 21%, respectively. Nonenhanced CT demonstrated > 20 Hounsfield Units (HU) for all tumours (median 34 (21-45)), median size 92 mm (20-196), Ki67 17% (1-40%), Weiss score 7 (4-9) and Helsinki score 24 (4-48). ACC was more often found in the left than the right adrenal (p < 0.05). One child had Beckwith-Wiedemann and one a TP53 mutation. In adults, the primary tumour was resected in 88 and 79% received adjuvant mitotane therapy. Median hospital stay was significantly shorter in the laparoscopic vs. open surgery group (4 (3-7) vs. 8 (5-38) days, respectively; p < 0.001). In 3/4 patients, prolonged remission of > 5 to > 10 years was achieved after repeated surgery of metastases. Overall 5-year survival was 67%, and 96% vs. 26% for ENSAT stage I-II vs. III-IV (p < 0.0001). ENSAT stage and Ki67 predicted survival, type of surgery did not. Mitotane associated with better survival. CONCLUSIONS: Contemporary ACC predominantly presents as an incidental imaging finding, characterised by HU > 20 on nonenhanced CT but variable tumour size (20-196 mm). Malignancy cannot be ruled out by small tumour size only. The 5-year survival of 96% in ENSAT stage I-III compares favourably to previous studies.

C-myc expression in adrenocortical tumours.

AIMS: Widespread use of high-resolution imaging techniques and thus increased prevalence of adrenal lesions has made diagnostics of adrenocortical tumours an increasingly important clinical issue. In non-metastatic tumours, diagnosis is based on histology. New or enhanced information for clinicopathological diagnosis, revealing the malignant potential of the tumour, could emerge by means of biomarkers. The connection of proto-oncogene c-myc to adrenocortical neoplasias is poorly known, although the Wnt/beta-catenin pathway, one of the signalling pathways leading to induction of c-myc expression, has been connected to development of adrenocortical neoplasias. We studied c-myc expression in adrenocortical tumours and investigated molecules associated with the signalling pathway of c-myc, including cell cycle-related proteins p27, cyclin E and cyclin D1. METHODS: We studied 195 consecutive adult patients with 197 primary adrenocortical tumours. Histopathological diagnosis was determined by Weiss score and the novel Helsinki score. C-myc, cyclin D1, cyclin E and p27 expressions were determined by immunohistochemistry. RESULTS: Benign adenomas showed prominent nuclear c-myc expression comparable to that of normal adrenocortical cells, whereas carcinomas showed increased cytoplasmic expression. Strong cytoplasmic and weak nuclear c-myc expressions associated with malignancy and adverse outcome. C-myc staining did not correlate with cyclin E. Cyclin D1 correlated with cytoplasmic c-myc expression and to a lesser extent with nuclear c-myc. P27 correlated with cytoplasmic c-myc, but not with nuclear c-myc. P27 correlated with cyclin E. CONCLUSIONS: Strong cytoplasmic c-myc expression and weak nuclear expression in adrenocortical tumours associated with malignancy and shorter survival.

Helsinki score-a novel model for prediction of metastases in adrenocortical carcinomas.

Histopathologic diagnosis of adrenocortical tumors is based on adverse features that indicate malignant potential. Proliferation index has served as a supplemental tool in assessing the malignant potential of adrenocortical tumors. None of the current histologic classification systems can sufficiently accurately predict tumors' metastatic potential. We studied 177 consecutive adult patients with primary adrenocortical tumors operated on at Helsinki University Central Hospital between 1990 and 2003, all patients with a minimum follow-up of 5 years. We determined for each tumor the Weiss score and the Weiss revisited score by Aubert. Proliferation index was measured by computer-assisted image analysis. Each of the 9 Weiss criteria and the proliferation index were then used to establish a scoring system to predict the metastatic potential of adrenocortical tumors. Use of stepwise regression analysis led us to propose a calculation: 3 × mitotic rate (>5/50 high-power fields) + 5 × presence of necrosis + proliferation index in the most proliferative area of the tumor. Using a cutoff value of 8.5, the new scoring system was able to diagnose metastatic adrenocortical carcinoma with 100% sensitivity (confidence interval [CI], 76.8%-100%) and 99.4% specificity (CI, 96.6%-100%). The corresponding sensitivity of the Weiss system was 100% (CI, 76.8%-100%), and specificity, 90.2% (CI, 84.6%-94.3%), with sensitivity of the Weiss revisited system at 100% (CI, 76.8%-100%) and specificity at 96.9% (CI, 93.0%-99.0%). The new Helsinki score thus was accurate in predicting the metastatic potential of adrenocortical tumors.

Adrenocortical tumours: high CT attenuation value correlates with eosinophilia but does not discriminate lipid-poor adenomas from malignancy.

BACKGROUND: Characterisation of adrenal tumours is an important clinical problem. Unenhanced CT is the primary imaging modality to assess the nature of these lesions. AIMS: To study the correlation between unenhanced CT attenuation value and the specific histopathology, as well as the proportion of lipid-poor eosinophilic cells in adrenocortical tumours. METHODS: We studied retrospectively primary adrenocortical tumours that had been operated on at Helsinki University Central Hospital between 2002 and 2008. Of 171 tumours, 79 had appropriate preoperative CT scans and were included in the study. We evaluated the unenhanced CT attenuation values (Hounsfield units, HU) of these tumours and determined their histopathological diagnosis by the Weiss scoring system. We also assessed the proportion of lipid-poor eosinophilic cells for each tumour. RESULTS: Unenhanced CT attenuation value (HU) in adrenocortical tumours correlated well with the proportion of lipid-poor eosinophilic cells (rs=0.750, p<0.001). HU and Weiss score also had a correlation (rs=0.582, p<0.001). CONCLUSIONS: Unenhanced CT attenuation value correlates well with the percentage of lipid-poor eosinophilic cells, but unenhanced CT attenuation value fails to differentiate between benign lipid-poor adenomas and malignant adrenocortical tumours. All adrenocortical tumours with unenhanced CT attenuation value ≤10 HU are histologically benign lipid-rich tumours.