In this work, we describe a dataset suitable for analyzing the extent to which hematoxylin–eosin (HE)-stained tissue contains information about the expression of Ki67 in immunohistochemistry staining. The dataset provides images of corresponding pairs of HE and Ki67 stainings and is complemented by algorithms for computing the Ki67 index. We introduce a dataset of high-resolution histological images of testicular seminoma tissue. The dataset comprises digitized histology slides from 77 conventional testicular seminoma patients, obtained via surgical resection. For each patient, two physically adjacent tissue sections are stained: one with hematoxylin and eosin, and one with Ki67 immunohistochemistry staining. This results in a total of 154 high-resolution images. The images are provided in PNG format, facilitating ease of use for image analysis compared to the original scanner output formats. Each image contains enough tissue to generate thousands of non-overlapping 224 × 224 pixel patches. This shows the potential to generate more than 50,000 pairs of patches, one with HE staining and a corresponding Ki67 patch that depicts a very similar part of the tissue. Finally, we present the results of applying a ResNet neural network for the classification of HE patches into categories according to their Ki67 label.

Hematoxylin and Eosin (HE) staining is the gold standard in histopathological examination of cancer tissue, representing the first step towards cancer diagnosis. The second step is a series of immunohistochemical stainings, including cell proliferation markers called the Ki67 index. Deep learning models offer promising solutions for improving medical diagnostics, while generative models provide additional explainability of predictive models that is essential for their adoption in clinical practice. Our previous work introduced a novel approach that utilises a conditional StyleGAN model for generating HE-stained images conditioned on the Ki67 index. This study proposes to employ this model for generating sequences of HE-stained images reflecting varying Ki67 index values. Sequences enable exploration of hidden relationships between HE and Ki67 staining and can enhance the explainability of predictive models, e.g., by generating counterfactual examples. While our previous research focused on assessing the quality of generated HE images, this study extends that work by evaluating the model’s ability to capture Ki67-related variations in HE-stained images. Additionally, expert pathologists evaluated generated sequences and proposed criteria for assessing their relevance. Our findings demonstrate the potential of the conditional StyleGAN model as part of an explainable framework for analysing and predicting immunohistochemical information from HE-stained images. Results highlight the relevance of generative models in histopathology and their potential applications in cancer progression analysis.

ALK detection was performed on 2813 EGFR-unmutated NSCLC cases by simultaneous use of immunohistochemistry (VENTANA® anti-ALK D5F3, Roche Molecular Systems, Inc., Rotkreuz, Switzerland) and fluorescence in situ hybridization with the ALK break apart and the ALK/EML4 fusion probe (ZytoVision, Bremerhaven, Germany). A total of 33 cases were positive discordant (FISH-positive, IHC-negative) and 17 cases were negative discordant (FISH-negative, IHC-positive). This study’s aim was to reevaluate the methods used and compare discordant samples to positive concordant samples in order to ellucidate the differences. FISH signal variants were examined and compared. Positive discordant cases featured one pattern of ALK rearrangement in 41.4%, two patterns in 48.3%, and three patterns in 10.3% of analysed samples, with a higher variability of detected patterns and a higher number of ALK copy gains. Positive concordant cases displayed one pattern of rearrangement in 82%, two patterns in 17.8%, and three patterns in 0.6% of analysed samples. The association between number of patterns and concordance/discordance was statistically significant (p < 0.05). Eleven positive discordant and two negative concordant cases underwent NGS analysis, which resulted in identification of ALK fusion in one positive discordant and two negative discordant cases. Positive protein expression regardless of FISH result correlated more with a positive NGS result compared to samples with a positive FISH result with negative protein expression. FISH analysis was able to detect atypical or heterogenous patterns of rearrangement in a proportion of cases with negative protein expression, which may be associated with more extensive genetic alterations rather than true ALK rearrangement.

Hematoxylin and Eosin (HE) staining is gold standard in histopathological examination of cancer tissue, representing first step towards cancer diagnosis. Second step is series of immunohistochemical stainings, including cell proliferation marker called Ki67 index. Deep learning models offer promising solutions for improving medical diagnostics, while generative models provide additional explainability of predictive models, which is essential for their adoption in clinical practice. Our previous work introduced novel approach that utilises conditional StyleGAN model for generating HE-stained images conditioned on Ki67 index. This study proposes to employ this model for generating sequences of HE-stained images reflecting varying Ki67 index values. Sequences enable exploration of hidden relationships between HE and Ki67 staining and can enhance explainability of predictive models, e.g., by generating counterfactual examples. While our previous research focused on assessing quality of generated HE images, this study extends that work by evaluating model’s ability to capture Ki67-related variations in HE-stained images. Additionally, expert pathologists evaluated generated sequences, proposing criteria for assessing their relevance. Our findings demonstrate potential of conditional StyleGAN model as part of explainable framework for analysing and predicting immunohistochemical information from HE-stained images. Results highlight relevance of generative models in histopathology and their potential applications in cancer progression analysis.

Introduction: Circulating tumor DNA is a promising noninvasive tool for tumor genotyping and recurrence monitoring in Colorectal Cancer (CRC), but its role in nonmetastatic disease remains unclear. This cross-sectional study evaluated preoperative plasma circulating cell-free DNA (cfDNA) in patients undergoing complete mesocolic excision (CME) with D3 lymphadenectomy. Methods: Preoperative plasma samples were collected from 32 patients with predominantly right-sided CRC; 31 with histologically confirmed malignancy were analysed. cfDNA was sequenced using the AVENIO ctDNA Expanded Panel targeting 77 cancer-related genes. Sequencing metrics, Variant Allele Frequency (VAF), variant types, and cfDNA variants were evaluated. Results: High-quality cfDNA was obtained from all patients (mean yield 0.69 ng/μL). Sequencing met performance criteria (>22 million read pairs, >86% mapped reads, depth 6779×/3523×). Variants were detected in 74% of the patients, most frequently in TP53 (n=11), APC (n=8) and KRAS (n=7). The mean VAF was 0.8%, with some variants exceeding 20%. The most frequent hotspot was KRAS c.38G>A (p.Gly13Asp) (n=4). Recurrent variants were observed in IDH2, MET, GNAS, CTNNB1, and PDCD1LG2, indicating molecular heterogeneity. Discussion: Preoperative tumor-naïve cfDNA sequencing enabled detection of established CRC driver alterations and rare variants, supporting its potential for minimally invasive molecular characterization. Compared with tumor-informed approaches, this strategy allows plasma profiling without prior tissue sequencing; however, variant interpretation remains limited without matched tumor and white blood cell DNA. Conclusion: Targeted cfDNA sequencing is feasible in preoperative CRC patients undergoing CME and provides a molecular baseline for future perioperative studies. Larger prospective cohorts with longitudinal follow-up are required to define its prognostic and clinical utility.