Search

Simple Summary: We developed an attention-based whole slide image (WSI)-level classification deep learning model employing surgically and endoscopically resected specimens to predict LNM in T1 CRC. Our AI model with H&E-stained WSIs and without annotations showed good performance power with the validation of an independent cohort in a single center. The area under the curve of our model was 0.781–0.824, higher than that of previous artificial intelligence (AI) studies with only WSIs. Our AI model, which showed the highest sensitivity (92.9%), reduced unnecessary additional surgery by 14.2% more than using the current JSCCR guidelines (68.3% vs. 82.5%). This revealed the feasibility of using an AI model with only H&E-stained WSIs to predict LNM in T1 CRC. According to the current guidelines, additional surgery is performed for endoscopically resected specimens of early colorectal cancer (CRC) with a high risk of lymph node metastasis (LNM). However, the rate of LNM is 2.1–25.0% in cases treated endoscopically followed by surgery, indicating a high rate of unnecessary surgeries. Therefore, this study aimed to develop an artificial intelligence (AI) model using H&E-stained whole slide images (WSIs) without handcrafted features employing surgically and endoscopically resected specimens to predict LNM in T1 CRC. To validate with an independent cohort, we developed a model with four versions comprising various combinations of training and test sets using H&E-stained WSIs from endoscopically (400 patients) and surgically resected specimens (881 patients): Version 1, Train and Test: surgical specimens; Version 2, Train and Test: endoscopic and surgically resected specimens; Version 3, Train: endoscopic and surgical specimens and Test: surgical specimens; Version 4, Train: endoscopic and surgical specimens and Test: endoscopic specimens. The area under the curve (AUC) of the receiver operating characteristic curve was used to determine the accuracy of the AI model for predicting LNM with a 5-fold cross-validation in the training set. Our AI model with H&E-stained WSIs and without annotations showed good performance power with the validation of an independent cohort in a single center. The AUC of our model was 0.758–0.830 in the training set and 0.781–0.824 in the test set, higher than that of previous AI studies with only WSI. Moreover, the AI model with Version 4, which showed the highest sensitivity (92.9%), reduced unnecessary additional surgery by 14.2% more than using the current guidelines (68.3% vs. 82.5%). This revealed the feasibility of using an AI model with only H&E-stained WSIs to predict LNM in T1 CRC. [ABSTRACT FROM AUTHOR]

Combined hepatocellular‐cholangiocarcinoma (cHCC‐CC) is a rare type of liver tumour that exhibits both hepatocytic and biliary differentiation within the same tumour. The histology and genomic alterations of recurrent/metastatic cHCC‐CC are poorly understood. We selected six patients with cHCC‐CC whose recurrent or metastatic tumours were histologically confirmed. Four patients with classic cHCC‐CCs and two with intermediate cell carcinomas (ICs) were included. The clinicopathological features were evaluated, and next‐generation sequencing was performed in 17 multiregional and longitudinal tumour samples. The histology of recurrent/metastatic lesions of classic cHCC‐CCs was variable: hepatocellular carcinoma (HCC) was observed in one (25.0%) patient, cHCC‐CC in one (25.0%) patient, and cholangiocarcinoma (CC) in two (50.0%) patients. Among 13 samples from four classic cHCC‐CC patients, the most frequent pathological variants were TP53 (46.2%), TERT promoter (38.5%), ARID1A mutations (23.1%), and MET amplification (30.8%). In the sequencing analysis of each HCC and CC component, three (75.0%) of the four classic cHCC‐CCs shared pathogenic variants. A large proportion of mutations, both pathogenic and those of undetermined significance, were shared by each HCC and CC component. Regarding ICs, the ATM mutation was detected in one patient. In conclusion, the histology of recurrent/metastatic cHCC‐CCs was heterogeneous. Genomic profiling of classic cHCC‐CCs revealed similar genomic alterations to those of HCC. Considerable overlapping genomic alterations in each HCC and CC component were observed, suggesting a monoclonal origin. Genetic alterations in ICs were different from those in either HCC or CC, suggesting the distinct nature of this tumour. [ABSTRACT FROM AUTHOR]

* Context.--The Ki-67 index is strongly prognostic and is used as a surrogate marker to distinguish luminal A from luminal B breast cancer types.Objective.--To investigate differences in Ki-67 index between core needle biopsy samples and matched surgical samples in breast cancer.Design.--We included patients with invasive breast cancer who did not receive neoadjuvant therapy. A total of 89 pairs of core needle biopsies and surgical specimens were collected, and the Ki-67 index was assessed in hot spot areas using an image analyzer. We applied a 14% Ki-67 index to define low versus high groups.Results.--The Ki-67 index was significantly higher in core needle biopsies than in surgical specimens (P < .001), with a median absolute difference of 3.5%. When we applied 14% as a cutoff, 16 of 89 cases (18%) showed discrepancy. Thirteen cases showed a high Ki-67 index in core needle biopsies but a low Ki-67 index in surgical samples. There were 10 cases (11.2%) that showed discordant luminal A/B types between core needle biopsy and the matched surgical specimen. The reasons for the discordance were poor staining of MIB1 accompanied by fixation issues and intratumoral heterogeneity of the Ki-67 index.Conclusions.--A significant difference in the Ki-67 index between core biopsy and surgical specimens was observed. Our findings indicate that it may be better to perform the Ki-67 assay on the core needle biopsy and the surgical specimen than on only one sample.(Arch Pathol Lab Med. 2018;142:364-368; doi: 10.5858/arpa.2017-0014-OA), Ki-67 is a biological tumor marker that reflects tumor proliferation. In breast cancer, the Ki-67 index is strongly prognostic (1) and supplies additional useful information. For the purpose of prognostication [...]

PIK3CA mutations in cancer regulate tumour immunogenicity. Given that PIK3CA mutation subtypes influence therapeutic responses to AKT inhibitor and that H1047R mutation confers selective growth advantages after immunotherapy, we hypothesised that immune phenotypes may depend on PIK3CA mutation subtypes. We investigated 133 gastric cancers (GCs) harbouring PIK3CA mutation [21 E542K (15.8%), 36 E545X (27.1%), 26 H1047X (19.5%), and 46 others (34.6%)]. Four patients (3.0%) had a combination of mutations (E542K + E545K in 3 patients and E545K + H1047R in 1 patient). Epstein–Barr virus (EBV) and microsatellite instability (MSI) status, PD‐L1 (programmed death‐ligand 1) combined positive score (CPS), and stromal tumour‐infiltrating lymphocytes (TILs) were assessed. Concurrent genomic alterations, GeoMx digital spatial profiling (DSP), and OPAL multiplex immunohistochemistry (mIHC) were analysed, and correlation between the two assays was investigated. Of the 133 PIK3CA‐mutant (PIK3CAm) GCs, MSI‐high GC was significantly frequent in the H1047X mutation subtype (p = 0.005), while EBV positivity did not affect the mutation subtypes. There was no significant survival difference between the E542K, E545X, and H1047X subgroups. However, in the subgroup analysis for EBV‐positive GC, H1047Xm GC showed a trend towards shorter survival than E542K and E545Xm GC (p = 0.090 and 0.062). With DSP analysis, H1047Xm GC showed elevated VISTA (p = 0.0003), granzyme B (p < 0.0001), CD4 (p = 0.0001), and CD45 (p < 0.0001) expression compared with the E542Km or E545Xm GC subgroups, and only VISTA expression remained significant (p < 0.0001) using OPAL mIHC. DSP and OPAL analyses showed a moderate correlation of CD4 (ρ = 0.42, p = 0.004) and CD8 (ρ = 0.62, p < 0.001) expression levels in a comparison of six antibodies. Immune‐related protein expression levels were evident when classified by the three PIK3CA hotspot mutations, and H1047Xm GC showed the highest immune‐related protein expression compared with E542Km or E545Xm GC. Our results demonstrated distinct immune profiles in GC with PIK3CA hotspot mutations using GeoMx DSP and OPAL mIHC, and there was a correlation between the two multiplex platforms. © 2023 The Authors. The Journal of Pathology published by John Wiley & Sons Ltd on behalf of The Pathological Society of Great Britain and Ireland. [ABSTRACT FROM AUTHOR]