Persistent gaps in access to quality diagnostic tests undermine progress toward improved health outcomes and resilience to disease outbreaks in Africa. Furthermore, resources for strengthening laboratory systems have faced growing constraints owing to recent reductions in official direct financial assistance, highlighting an urgent need to identify investments that are high impact and minimally donor reliant. In this Perspective, we present a set of priority systems strengthening interventions based on expected impact and feasibility. These include the need for efficient and integrated testing networks, the establishment and implementation of national essential diagnostic lists, modernized procurement and supply chain practices, improved digital health standards and targeted strengthening of testing infrastructure for epidemic-prone diseases. These initiatives should be customized to local contexts and supported by reliable financing, enhanced national leadership and management capacity and updated policies. We propose that collective action focused on these priorities will improve health outcomes and be cost-saving-and will be superior to existing, more fragmented efforts to close gaps in testing. This also provides a pathway to self-reliance in health security and universal healthcare in the Africa region.

Background Artificial intelligence for digital pathology may improve cancer detection and workflow efficiency in low- and middle-income countries, but most models are trained and validated in high-income settings, creating uncertainty about generalizability and real-world deployability under differing pre-analytic and infrastructure conditions. Objective To validate a locally optimized colorectal carcinoma screening pipeline, retrained on data from a local Kenyan cohort, and to assess its feasibility as a sensitivity-forward assistive pre-screening workflow. Methods Formalin-fixed, paraffin-embedded hematoxylin and eosin-stained slides from 136 biopsy-proven colonic adenocarcinoma cases and 20 normal controls from 2 Kenyan institutions were digitized at 40× using a Grundium Ocus scanner. Whole-slide images were partitioned at the case level. A feature-enrichment and candidate tile-selection step identified adenocarcinoma-rich regions. Candidate 512 × 512 RGB tiles were then adjudicated by a gastrointestinal pathologist into adenocarcinoma-containing and benign/non-neoplastic tile libraries. These expert-curated tiles were used to train and validate a supervised convolutional neural network classifier, which generated tile-level malignancy probabilities. Tile scores were aggregated into case-level predictions using Top-K pooling and positive-tile burden rules to support sensitivity-forward screening. Results Among 14,452 tiles from 70 quality-controlled cases and 19 controls, the model achieved strong tile-level discrimination (sensitivity 0.9548, specificity 0.9926, F1 score 0.9686, and AuROC curve 0.966); case-level Top-K aggregation achieved an AuROC of 1.0. Conclusion A locally optimized computational pathology screening pipeline, based on data from a local population, demonstrated robust colorectal adenocarcinoma detection in a Kenyan cohort. These findings suggest that existing machine-learning models can be adapted to local populations through retraining with locally derived data.

AIM:To understand whether the worldwide implementation of PD-L1 testing in triple-negative breast cancer (TNBC) can be achieved in routine clinical practice. METHODS AND RESULTS:The multicentre retrospective observational VANESSA study consecutively and uniformly enrolled patients treated with systemic therapy for early or metastatic (e/m)TNBC diagnosed between 2014 and 2017. PD-L1 status was retrospectively assessed locally and centrally using the VENTANA PD-L1 (SP142) Assay (PD-L1 expression on tumour-infiltrating immune cells covering ≥1% of the tumour area). The primary objective was to determine the prevalence of PD-L1 positivity assessed locally on primary and/or metastatic tumour tissue. Concordance between local and central testing was a secondary endpoint. PD-L1-positive prevalence was 38% in eTNBC (728/1902) and 20% in mTNBC (30/152) and was higher in submitted tissue size >5 versus <5 mm diameter (eTNBC: 43% versus 16%; mTNBC: 24% versus 13%). Among 1967 samples tested both centrally and locally, concordance was 75% (Cohen's κ coefficient 0.52, 95% CI 0.48-0.55) and was similar regardless of cohort (eTNBC versus mTNBC), sample collection method (biopsy versus resection) or sample origin (primary versus metastatic). PD-L1-positive prevalence was higher by central versus local assessment (eTNBC: 55% versus 39%; mTNBC: 26% versus 20%). CONCLUSION:In this real-world study, PD-L1-positive prevalence was lower than in prospective trials assessing PD-L1 status centrally, lower in mTNBC than eTNBC, lower in smaller than larger tissue samples and lower by local than central assessment. These findings underline the importance of central PD-L1 testing on sufficiently large samples to ensure optimal selection for therapies targeting PD-(L)1 in mTNBC.

Abstract Background Population-based cancer registries remain limited in low- and middle-income countries (LMICs). In Kenya, their coverage is below the 20% threshold recommended by the International Agency for Research on Cancer. Pathology-based registries can provide complementary data in such contexts. This study describes the distribution of cancers diagnosed at a tertiary referral hospital in Nairobi over ten years, and evaluates the completeness of pathology reporting for selected cancers using College of American Pathologists (CAP) protocols. Methods We conducted a retrospective review of all histologically confirmed cancers at the Aga Khan University Hospital, Nairobi, from 2015 to 2024. Data were extracted from the institutional database. The data included region of origin, demographics, cancer site, and histologic diagnosis. Completeness of reporting was assessed in stratified random samples of breast and colorectal cancer reports against the College of American Pathologists (CAP) cancer reporting protocols. Results A total of 32,445 cancer cases were diagnosed over the study period, including 18,899 females and 13,546 males (58.3% and 41.7%, respectively). The median age at diagnosis was 52 years (Interquartile range 41–64) in females and 61 years (Interquartile range 47–71) in males, with an overall median age of 56 years (Interquartile range 43–68). Most cases occurred in the decades between 51 and 60 years (6816/32,445) and 61–70 years (6556/32,445). Overall, breast cancer was the most frequent malignancy (7236/32,445; 22.3%), followed by esophageal (3839/32,445; 11.8%) and prostate cancer (3023/32,445; 9.3%). Among women, breast (6926/18,899; 36.6%), cervical (2230/18,899; 11.8%), and esophageal cancers (1610/18,899; 8.5%) predominated, while among men, prostate (3023/13,546; 22.3%), esophageal (2229/13,546; 16.5%), and non-Hodgkin lymphoma (957/13,546; 7.1%) were most common. Pediatric cancers (≤ 19 years) accounted for 1325 cases (4.1%). Completeness of pathology reporting exceeded 95% for most required elements, and it improved from 83.3% in 2015 to over 98% in 2024. Conclusion This large pathology-based dataset highlights the cancer burden in Kenya, with breast and prostate cancers predominating among females and males, respectively. There are regional disparities in sources of referral for certain cancer types. Pediatric malignancies represented a small proportion of cases. Pathology reporting for key data elements was generally of high completeness.

Abstract Background: The tumor immune microenvironment (TIME) reflects both tumor-intrinsic biology and host-modifiable factors. Characterizing how tumor features, body mass index (BMI), and reproductive history relate to immune activity may help explain heterogeneity in breast cancer and inform precision prevention and survivorship strategies. Methods: We analyzed 461 invasive breast tumors from Kenyan patients using a custom NanoString nCounter® immune-focused gene panel. Intrinsic subtypes were assigned with PAM50, and immune cell composition (relative proportions for 22 immune cells) was estimated by CIBERSORTx. Composite immune activity scores were derived to represent functional modules: z_hot (CD8+, M1 macrophages, NK activated, T follicular helper) and z_suppression (Treg + M2 macrophages) from CIBERSORTx; z_cytotoxic (GZMB, PRF1), z_exhaustion (PDCD1, LAG3, CTLA4), and z_checkpoint (PDCD1, PDCD1LG2, CTLA4, LAG3) from GSVA signatures. Associations between immune scores and tumor features (PAM50 subtypes, risk of recurrence (ROR), RNA-based TP53 status, tumor grade) or host factors (BMI, menopausal status, parity, breastfeeding) were evaluated in age-adjusted linear regression models. Independent effects of host factors were further examined in multivariable regression models adjusted for tumor characteristics and lifestyle covariates. Multiple testing was accommodated using a false discovery rate adjusted p-values. Results: Participants had a mean age of 50.3 years; 43.8% of tumors were luminal A and 21.5% basal-like. Most women were overweight/obese (BMI ≥ 25, 73.1%) and 65.7% had ≥ 3 children. High-grade, basal-like, high-ROR, and TP53 mutant-like tumors showed strong evidence of elevated z_hot, z_cytotoxic, z_checkpoint, and z_exhaustion scores (BH-adjusted p < 0.001), along with modestly lower z_suppression in basal-like and TP53 mutant-like tumors (p < 0.05), indicating an active TIME characteristic of aggressive tumors. In contrast, established BC risk factors showed weaker influence on TIME. Overweight/obese patients were more likely to have cold TIME (lower CD8 (p = 0.09), T follicular helper (p < 0.05), and z_hot score (p < 0.05)), while patients with longer duration of breastfeeding had more active TIME (higher T follicular helper and z_cytotoxicity; lower M2 macrophages and z_suppression (all p < 0.05)). Conclusions: Tumor-intrinsic subtype classification remains the dominant determinant of immune heterogeneity, while established BC risk factors (BMI and reproductive risk factors) may modulate immune responsiveness. Integrating tumor, reproductive, and lifestyle data can help clarify immune variation across diverse populations. Citation Format: Li Feng, Amber N. Hurson, Shahin Sayed, Hela Koka, Viviane Oluoch, Veronica Ngundo, Alfred Mburu Githuka, Zaitun Ajuoga, Shaoqi Fan, Kristine Jones, Belynda Hicks, Amy Hutchinson, Maria Brown, Petra Lenz, Aaron M. Rozeboom, Difei Wang, Francis Makokha, Stefan Ambs, Jonine D. Figueroa, Ruth M. Pfeiffer, Xiaohong Rose Yang. Tumor and Host Determinants of the Breast Tumor Immune Microenvironment in Kenyan Breast Cancer Women [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 1 (Regular Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(7 Suppl):Abstract nr 7606.

Despite growing evidence on the impact of intra-tumoral heterogeneity (ITH) in breast cancer (BC), its biological drivers and clinically relevant mitigation strategies remain poorly characterized, particularly in low-resource settings where spatial profiling is rarely applied. To address this gap, we conducted a comprehensive, multiplatform spatial profiling study of BC in a Kenyan population. Using the NanoString GeoMx™ Digital Spatial Profiler, we quantified 44 protein markers across 707 spatially defined, PanCK-segmented Areas of Illumination (AOIs) from 31 tumors. All tumors exhibited measurable ITH, with one-third showing marked heterogeneity. Sixteen patients exhibited two or more intrinsic subtypes within their tumors, three of whom had three subtypes. Immune-related ITH was particularly pronounced for CD8 and CD68. AOI phenotype (epithelium- vs. stroma-rich) emerged as the primary driver of marker variability. Linear mixed-effects (LME) modeling further revealed that intrinsic subtype influenced not only classical subtyping markers (ER, PR, HER2) but also immune and signaling markers including IDO1, S100B, PTEN, and BCL-2. Tissue morphology and spatial neighborhood contributed additional variance, particularly in stroma-rich AOIs, while incorporating AOI spatial coordinates improved model fit and revealed further spatially structured heterogeneity. Nonetheless, over 25

Cancer now ranks among the leading causes of death in Kenya, with approximately 35,867 new cases and 22,888 deaths annually, and outcomes remain constrained less by the availability of therapy than by the diagnostic pathway that must precede it. A national pathology workforce numbering in the low hundreds serves a population exceeding fifty million, and delayed or absent tissue diagnosis drives late presentation across the five malignancies that account for most of the national burden. Artificial intelligence offers a mechanism to extend interpretive capacity in a system where human expertise cannot be scaled at the rate the disease burden demands. This communication sets out a prioritized research agenda spanning smartphone-assisted visual inspection of the cervix, breast ultrasound classification and histopathological grading, morphological and immunophenotypic triage of hematolymphoid neoplasms, radiomic triage of thoracic and hepatic disease, and AI-assisted radiotherapy contouring. We present a six-stage methodological pipeline covering stakeholder engagement, local dataset curation, transfer learning, validation, usability assessment, and data governance, together with a tiered adoption framework matched to laboratory capability and the workforce competencies required for safe clinical supervision of these systems. The agenda is directed at early-career investigators and is transferable to comparable low- and middle-income settings.

Digital pathology (DP) represents a major advance in global health, integrating digital whole-slide imaging, telepathology and clinical-grade artificial intelligence into cohesive workflows that can improve diagnostic accuracy and educational access. Realising these gains, however, depends on local infrastructure, governance, reimbursement and workforce capacity. The COVID-19 pandemic accelerated DP uptake in some settings, but adoption remains highly uneven between high-resource settings (HRS) and low-resource settings (LRS), reflecting unequal infrastructure, workforce, policy and financing. This analysis contends that reciprocal innovation, a model of ongoing two-way learning between HRS and LRS that is distinct from one-way technology transfer, can accelerate DP adoption by leveraging the strengths of each environment. Drawing on examples from HRS and LRS, we examine use cases in education and clinical practice, the adoption barriers that persist and policy mechanisms required for sustainable scaling. While HRS contribute regulatory frameworks, validation standards and mentorship, LRS offer innovations in cost efficiency, workflow resilience and open-source adaptation that can inform more sustainable models globally. Neither HRS nor LRS are homogeneous entities. Resource availability also varies substantially within each category: HRS may contain poorly resourced institutions, while some institutions in LRS may have comparatively greater resources but still face broader system-level infrastructure constraints. Ensuring access to those in LRS requires targeted investment, harmonised standards and mutual learning across resource levels. DP has the potential to foster bidirectional partnerships to democratise access to timely, accurate and patient-centred pathology services across all resource settings.

Breast cancer biology is influenced by both genetic ancestry and the environment with implications for disease outcomes. However, we lack the knowledge how non-cancerous cells within tumors are functionally altered by these factors at single-cell resolution. Here, we created a single-nucleus (sn) dataset from 33 African American, 25 Kenyan, and 24 European American women with breast cancer using combined RNA and ATAC sequencing. We successfully isolated intact, high-quality single nuclei from archival frozen breast tumor tissue using an optimized combination of enzymatic digestion and automated tissue homogenization. SnMultiome sequencing of 82 tumors was performed using the 10x Genomics platform. Following filtering, normalization, peak calling, and integration, our dataset includes a total of 292,458 nuclei. Cancerous (163,419 nuclei) and non-cancerous (129,039 nuclei) cells were distinguished based on DNA copy number. Within the microenvironment, 11 major immune, epithelial, and stromal cell types were identified. CD45+, EPCAM- cells were further analytically extracted, clustered, and manually annotated based on known marker genes, detecting 26 distinct immune subpopulations across our samples. After adjustment for potential confounders (age, body mass index, and tumor molecular subtype) in all analyses, we found that women of African descent possessed distinct, activated mesenchymal and endothelial cell subpopulations within the tumor vasculature that promote inflammation, angiogenesis, and cancer cell invasion. Women of African descent exhibited a more immune-suppressive tumor microenvironment, with elevated expression of immune checkpoint markers (PD-1, TIGIT, CTLA-4, and LAG-3), functional impairment of dendritic cells, and enrichment of regulatory and tolerogenic lymphocytic subpopulations. Further, malignant cells from African descent patients expressed a discrete set of CAR-T targets with potential implications for precision immunotherapy. Lastly, we developed three gene signatures within the microenvironment that are predictive of survival with external validation in both tumor-adjacent normal and cancerous tissues from the TCGA. Together, we uncovered molecular breast tumor characteristics of clinical significance in women of African ancestry. Alexandra R. Harris, Huaitian Liu, Brittany D. Jenkins, Sanna Madan, Tiangen Chang, Tiffany H. Dorsey, Eytan Ruppin, Shahin Sayed, Francis Makokha, Gretchen L. Gierach, Stefan Ambs. Single-nucleus sequencing reveals a functionally distinct breast tumor microenvironment by ancestral group with implications for disease progression and therapy [abstract]. In: Proceedings of the 18th AACR Conference on the Science of Cancer Health Disparities; 2025 Sep 18-21; Baltimore, MD. Philadelphia (PA): AACR; Cancer Epidemiol Biomarkers Prev 2025;34(9 Suppl):Abstract nr C109.

Tumor-infiltrating immune cells in the tumor microenvironment (TME) are critical for breast cancer (BC) progression and treatment outcomes, with notable variations observed across populations. Unique BC risk factors, such as germline genetics, mammographic density, infectious exposures, and lifestyle factors, may shape distinctive immune profiles and lead to distinct tumor characteristics and clinical outcomes in different populations. Comparing immune cell compositions across populations can improve our understanding of BC racial disparity and guide personalized treatment strategies. This study aims to characterize and compare immune composition in the BC TME between East Africa (Kenya) and East Asia (Hong Kong) when accounting for age, intrinsic subtype, and analytical methods. We collected formalin-fixed paraffin-embedded (FFPE) tumor blocks from BC patients in Kenya and Hong Kong, processed them for nucleic acid extraction and analyzed them using NanoString nCounter from a custom-designed code set including 50 PAM50 genes and 547 CIBERSORT panel genes. Samples were selected to match age and estrogen receptor (ER) status distributions. After removing duplicates and QC-failed samples, 467 samples from Kenya (mean age=50.57, ER+: 71.18%) and 118 samples from Hong Kong (mean age=54.94, ER+: 72.03%) were included in the subsequent analyses. CIBERSORTx was used to estimate relative fractions of 22 immune cell types. PAM50 subtype was defined as Luminal A, Luminal B, HER2-enriched, Basal-like, or Normal-like using the nearest centroid method. Multiple linear regression models were performed to identify the differences in immune cell proportions by ethnicity, adjusting for age, PAM50 subtype, and tumor grade. After filtering cells with mean fractions <5%, this analysis was focused on T follicular helper, CD8+, resting CD4+, activated NK, resting mast, and macrophage subsets (M0, M1, M2). Compared to Hong Kong, Kenyan samples exhibited higher relative fractions of CD8+ T cells (p < 0.01), activated NK cells (p < 0.001), and macrophages M2 (p < 0.05) but lower fractions of resting CD4+ T cells (p < 0.05) and macrophages M0 (p < 0.01). Similar results were observed with absolute abundance analyses. This study reveals distinct immune profiles between Kenyan and Hong Kong BC TMEs. Kenyan samples exhibit a relatively “hot” immune TME with both pro-inflammatory and immunosuppressive components, while Hong Kong samples show a more quiescent or relatively “cold” immune environment. These population-specific differences highlight the potential need for tailored immunotherapy strategies considering unique immune characteristics shaped by genetic and environmental factors. Future analyses will explore associations with clinical outcomes and risk factors to elucidate mechanisms driving these racial disparities. Li Feng, Shahin Sayed, Hela Koka, Viviane Oluoch, Veronica Ngundo, Alfred M. Githuka, Zaitun Ajuoga, Kristine Jones, Belynda Hicks, Amy Hutchinson, Maria Brown, Petra Lenz, Aaron M. Rozeboom, Difei Wang, Francis Makokha, Stefan Ambs, Jonine Figueroa, Xiaohong R. Yang. A comparison of tumor microenvironment immune profiles using CIBERSORTx in breast cancer patients from Kenya and Hong Kong [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 5358.

Chimeric antigen receptor (CAR) T-cell therapy has revolutionized the treatment of hematological malignancies. However, its application in solid tumors remains limited because single targets are unlikely to suffice due to tumor antigen heterogeneity and off-tumor toxicities. To overcome these obstacles, we developed LogiCAR designer, a computational approach that utilizes single-cell transcriptomics data from patient tumors to systematically identify the cancer-specific antigen circuits with logic gates ("AND," "OR," and "NOT") that target the majority of cancer cells in a tumor while sparing normal cells and tissues as much as possible. LogiCAR designer efficiently scales to higher-order antigen combinations involving up to five genes. Applied to a large-scale dataset encompassing approximately 2 million cells (including > 620k tumor cells) from 342 clinical patient samples across all major breast cancer subtypes, LogiCAR designer identified antigen circuits with enhanced tumor-targeting efficacy and improved safety profiles compared to both previously reported circuits and single-target therapies in clinical trials. However, even these optimized shared circuits still proved insufficient for some patients. We hence systematically studied LogiCAR designer's ability to identify highly effective CAR circuits that are individualized to each patient. Remarkably, such personalized CAR circuits provide estimated tumor-targeting efficacy tantamount to complete response in 76% of patients and partial response for all patients. Taken together, this analysis is the first systematic quantification of the efficacy and safety of all possible CAR circuits, showing that: (a) the quality of existing solutions leaves much to be desired; (b) the ability of shared circuits optimized across many patients is moderate, and finally, (c) individually tailored circuits offer significantly higher tumor-targeting efficacies for patients. LogiCAR designer offers a rigorous, data-driven way to facilitate the rational design of safe and effective CAR-based immunotherapies for cancer.

We present a review of mature B-cell neoplasms as described in the fifth edition of the WHO classification of haematolymphoid tumours (WHO-HAEM5). Entities have expanded, and definitions are increasingly reliant on genomic and other technologies. However, the WHO-HAEM5 employs a hierarchical structure with family (class)-level definitions that group several specific entities. This approach enables the assignment of a family-level diagnosis when criteria for specific entities cannot be met due to resource constraints. To facilitate application in resource-limited settings, WHO-HAEM5 divides diagnostic criteria into 'essential' and desirable criteria for most entities. This review focuses on changes and updates in B-cell lymphoma classification, providing guidance on how to apply the WHO classification in resource-limited settings.

Accounting for over 25% of cancer incidence, breast cancer is the second most common cause of cancer mortality among females in Kenya. Over 80% of Kenyan women diagnosed with breast cancer are unlikely to survive more than 5 years. Multiple barriers exist to early diagnosis, including access to specialist services and affordability of ancillary testing such as immunohistochemistry (IHC). The aims of our study were to evaluate the performance of IHC for estrogen (ER), progesterone (PR), HER2, and Ki67 in subtyping of breast cancer tissue by comparing them against the gold standard PAM50 intrinsic subtypes and to establish an optimal strategy for biomarker testing in a resource-limited setting such as Kenya. A total of 467 formalin-fixed paraffin-embedded (FFPE) breast tumor tissues from newly diagnosed Kenyan women with breast cancer were analyzed using NanoString nCounter (fifty PAM50 genes within a ∼ 600 gene panel) for transcriptome-based subtyping according to PAM50 classification. IHC-based subtypes were assigned to the same tumors following the ASCO/CAP guidelines for ER/PR/HER2 and Ki67 when available (with a cutoff of >/= 20%) and compared to PAM50 subtypes using Kappa statistics as well as subtype-specific sensitivity and specificity. Overall, the cohort mean age was 50.57 years (SD = 12.44), with 50% of tumors classified as grade 3 and 42% at advanced stages (III/IV). The intrinsic subtype classification by PAM50 categorized 205 (43.9%), 83 (17.8%), 51 (10.9%), 99 (21.2%), and 29 (6.2%) as Luminal A, Luminal B, HER2-enriched, basal-like, and normal-like, respectively. The highest concordance between IHC-based and PAM50 transcriptome-based subtyping was observed when Ki67 was added to distinguish Luminal A from Luminal B. This improvement with Ki67 was sustained even when PR was excluded. The proportion of Luminal B subtypes increased by 28.41%, with 77 additional cases reclassified from Luminal A to Luminal B out of 271 Luminal patients, incorporating the proliferation marker Ki-67, or Grade, as a surrogate if Ki-67 was unavailable. Overall, there was a good concordance of IHC and PAM50 for Luminal A and Basal-like breast cancer based on ER, PR, Her2. The incorporation of Ki67 and/or grade can improve the diagnosis of Luminal B patients who may benefit from therapy in addition to endocrine therapy. We note that PR could potentially be replaced by Ki67 where resources are limited. Shahin Sayed, Li Feng, Hela Koka, Difei Wang, Maria Brown, Kristine Jones, Margaret Maina, Veronica Ngundo, Alfred M. Githuka, Zaitun Ajuoga, Shaoqi Fan, Amber Hurson, Xing Hua, Petra Lenz, Rosemary W. Kamau, Beryl A. Ooro, Francis Makokha, Stefan Ambs, Aaron M. Rozeboom, Amy Hutchinson, Belynda Hicks, Mustapha Abubakar, Jonine Figueroa, Xiaohong R. Yang. Optimizing Breast Cancer Subtyping in Kenya: Concordance Between Immunohistochemistry and PAM50 Molecular Classification and Strategies for Resource-Limited Settings [abstract]. In: Proceedings of the 13th Annual Symposium on Global Cancer Research; 2025 Sep 16. Philadelphia (PA): AACR; Cancer Epidemiol Biomarkers Prev 2025;34(12_Suppl):Abstract nr 61.

The 5th edition of the World Health Organization Classification of Tumours (WCT) serves as a foundation for global diagnostic standards in tumour pathology. Similar to other volumes in this series, the Skin Tumours (Skin5) edition follows a standardized approach. This edition introduces two new chapters: 'Tumours of the nail unit' and 'Metastases to skin', along with new entities across relevant chapters. This review article provides an overview of the updates in Skin5 based on currently published evidence, with emphasis on newly introduced chapters and newly described entities that involve the skin, in particular, epidermal, melanocytic and appendageal tumours.

The Lancet Commission on diagnostics made recommendations for ten topics: national strategy (including national essential diagnostics lists), access in primary care, workforce, regulatory framework, national financing, affordability, appropriate use of technology, needs in conflict or fragile situations, advocacy, and an international alliance with oversight capabilities. Since 2021, progress in these areas has benefitted greatly from the adoption of a World Health Assembly resolution on diagnostics and the work of a broad coalition, as assessed by literature surveys by subject matter experts, quantitative findings (where feasible), and an anonymous survey of knowledgeable and engaged individuals. Greater progress was observed where there was political will and the production of diagnostics coincided with industrial policy goals, also in areas where changing the legal and health policy frameworks was involved. Progress was slower on recommendations with substantial resource implications (eg, labour force, affordability, and diagnostics for conflict situations). It is expected that the Global Diagnostics Coalition will consolidate and accelerate progress.

Breast cancer is a highly heterogeneous disease at the clinical, population, and molecular levels. Using key molecular markers, breast cancer can be classified into at least four major intrinsic subtypes: luminal A, luminal B, HER2-enriched, and basal-like or triple-negative. These subtypes are associated with distinct risk factors, clinical behaviors, treatment responses, and outcomes. Recent research highlights the importance of the breast tumor microenvironment (TME) in shaping tumor evolution and cancer progression. Both tumor and TME cells within an individual patient often exhibit substantial intratumoral heterogeneity (ITH), the extent of which may impact accurate tumor classification and treatment response, but has not been well characterized especially in women of African Ancestry. The goal of this study was to comprehensively investigate ITH of intrinsic subtypes and immune TME markers in breast cancer patients from an East African population. Specifically, we employed the Nanostring GeoMxTM platform to quantify 44 key breast cancer intrinsic and immune subtyping markers in spatially defined regions of breast tumors from 31 Kenyan breast cancer patients. We found that all 31 tumors exhibited some degree of ITH, with a third of them showing substantial ITH. Cell composition (epithelial vs. stromal), spatial location of TME immune cells (tumor-leading edge, proximity to tumor or stroma), and intrinsic subtype (specifically, basal-like vs. luminal A) all contributed significantly to ITH of these markers. When comparing results to those obtained from an in-situ multiplex staining platform InSituPlex (ultivue) and bulk Nanostring nCounter RNA expression, all generated from the same set of cases, we found that intrinsic or immune subtyping by bulk approaches, which is heavily influenced by the predominant subtypes within a tumor, may lead to inaccurate classification when substantial ITH exists. Similarly, sampling-based approaches like tissue microarrays (TMAs) or region-of-interest (ROI)-based assays are also likely to be subjective to inaccurate and inconsistent measurements of immune markers due to sampling bias. In low-resource settings, in-situ assays combined with multi-region scoring may offer a more accurate yet practical alternative to bulk or single-cell assays, which are often costly and technically challenging. Mustapha Abubakar, Shahin Sayed, Hela Koka, Scott Lawrence, Karun Mutreja, Xing Hua, Viviane Oluoch, Veronica Ngundo, Zaitun Ajuoga, Difei Wang, Chad Highfill, Petra Lenz, Maria Brown, Aaron M. Rozeboom, Kristine Jones, Amy Hutchinson, Belynda Hicks, Paul S. Albert, Francis Makokha, Stefan Ambs, Jonine Figueroa, Jianxin Shi, Xiaohong Rose Yang. Multi-platform spatial profiling reveals intra-tumor heterogeneity in intrinsic subtypes and immune markers in breast tumors from Kenyan patients [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 5316.

Current CAR-T therapies for solid tumors are limited by suboptimal target selection, with single antigens failing to address tumor heterogeneity and causing on-target, off-tumor toxicities. Systematic identification of multi-target combinations using logic gates (AND, OR, NOT) represents an untapped approach for discovering safer, more effective CAR immunotherapy targets. We developed LogiCAR designer, a genetic algorithm-based framework that systematically screens 2,758 cell surface proteins to identify optimal 1-5 gene logic-gated target combinations (i.e., ‘circuits’) from single-cell transcriptomics data. Applied to the largest breast cancer single-cell dataset (∼2 million cells, >620k tumor cells from 342 patients across 17 cohorts), we optimized tumor-targeting efficacy while maintaining safety across 689,601 normal cells from 31 Human Protein Atlas tissues. Comprehensive target validation included RNA and protein expression profiling across major human tissues and tumor microenvironment specificity analysis. Our systematic approach identified novel cell surface targets with superior profiles compared to current clinical candidates. The top 3-gene circuit ('GABRP | PRLR | VTCN1') achieved 60% tumor-targeting efficacy—234% higher than the best clinical trial targets. Newly discovered single targets (ELAPOR1, PRLR, BAMBI, LDLRAD3) demonstrated consistent tumor-specific expression (>2-fold tumor vs. non-tumor cells, p<0.05) across all datasets, unlike current clinical targets. Safety profiling revealed that many existing targets (BSG, EPCAM, TACSTD2) showed concerning expression in critical normal tissues, while our identified circuits maintained superior safety profiles. Finally, shared circuits were still ineffective for some patients, prompting development of personalized approaches. Individualized target combinations addressed intratumor heterogeneity: in our new 82-patient multi-ethnicity cohort, personalized circuits reached 98% mean efficacy with 76% of patients achieving complete response-equivalent targeting. This work establishes the first systematic molecular target discovery pipeline for multi-antigen therapeutic design, identifying previously unrecognized targets with superior efficacy-safety profiles. Our approach transforms target selection from empirical to data-driven, providing a foundation for next-generation precision therapeutics across cancer types. Sanna Madan, Tian-Gen Chang, Andrew Martinez, Alexandra R. Harris, Huaitian Liu, Saugato R. Dhruba, Binbin Wang, Padma S. Rajagopal, Sanju Sinha, Aravind Srinivasan, Simon R. V. Knott, Shahin Sayed, Francis Makokha, Chi-Ping Day, Gretchen L. Gierach, Stefan Ambs, Alejandro A. Schäffer, Eytan Ruppin. Systematic discovery of logic-gated cell surface targets for enhanced solid tumor CAR therapy through single-cell transcriptomics [abstract]. In: Proceedings of the AACR-NCI-EORTC International Conference on Molecular Targets and Cancer Therapeutics; 2025 Oct 22-26; Boston, MA. Philadelphia (PA): AACR; Mol Cancer Ther 2025;24(10 Suppl):Abstract nr C042.

INTRODUCTION:Targeted therapies for patients with non-small cell lung cancer (NSCLC) have significantly improved the outcomes for patients with oncogenic driver mutations. Testing for epidermal growth factor receptor (EGFR) mutations, anaplastic lymphoma kinase (ALK) rearrangement, c-ros oncogene 1 (ROS1) rearrangement and programmed death ligand 1 (PDL 1) testing are now highly recommended for all patients with NSCLC. There are clear geographical differences in the rate of driver mutations with higher EGFR mutation rates in the Asian population in comparison to the Caucasian population. Little is known about the rate of oncogenic driver mutations in predominantly Black African populations like those living in the East African region. This study describes the oncogenic driver mutations found in lung cancer patients diagnosed in a single institution in Kenya. PATIENTS AND METHODS:We retrospectively reviewed the charts of patients who had a pathological diagnosis of NSCLC at Aga Khan University Hospital, Nairobi (AKUHN) between January 2012 and December 2022. Data was analyzed for socio-demographic characteristics, smoking status, clinical stage, histological sub-type and presence or absence of driver mutations. RESULTS:123 cases of non-small cell lung cancer (NSCLC) were included in the analysis. The median age at diagnosis was 62 years (IQR: 53.0 - 71.0) and 41.5 % (n = 51) of patients were under 60 years at time of diagnosis, 47.2 % (n = 58) were female and 78.9 % (n = 97) were of Black African descent. Only 29.4 % (n = 30) were known smokers whereas 73.2 % (n = 90) had stage IV disease. Adenocarcinoma was the most common sub-type in 85.4 % (n = 105) patients and 60 % of cases had EGFR testing done and a mutation was detected in 35 % (n = 26/74) patients tested. ALK and ROS rearrangement was positive in 19.5 % (n = 8/41) and 6.0 % (n = 2/33) patients tested respectively. Only 23.5 % (n = 8/34) of patients tested for PDL1 had an expression of >1 %. There was no significant association between gender, ethnicity and smoking with EGFR mutation. CONCLUSION:In our setting, EGFR testing was the most common molecular test done for lung cancer with a positive rate like that reported in Asian communities. Further studies are needed in a larger population to define further the molecular profile of lung cancer in Sub Saharan Africa. Access to testing will enhance targeted therapies for patients leading to improved outcomes.