Robert Lukande

Introduction: Cancer incidence is rising rapidly in low- and middle-income countries (LMICs), where diagnostic pathology capacity is constrained. AI applied to routine histopathology could strengthen diagnostic capacity and offer leapfrogging potential by embedding AI into diagnostic systems as they are built up. However, without empirically identifying implementation determinants in LMIC cancer-diagnostic settings, AI solutions risk being poorly aligned with local constraints, limiting impact or exacerbating existing disparities. Methods: We conducted an exploratory mixed-methods pilot study across three phases: 1) a structured survey characterising the diagnostic and technical landscape and AI priorities; 2) focus group discussions (FGDs) to identify implementation determinants and 3) a second survey quantifying FGD-derived themes. Twelve physicians (nine pathologists, three gynaecologists) were purposively recruited (10 LMIC, 2 high-income countries for reference) and findings integrated through narrative triangulation. Results: Diagnostic and AI infrastructure varied substantially and did not track consistently with country income classification. H&E staining was universally available, immunohistochemistry capacity was below WHO-recommended capacity for cancer diagnosis at 7 of 10 sites and molecular testing absent or minimal at 6 of 10 sites. Whole-slide-imaging scanners were available at 8 of 10 sites, graphics processing units, required for efficient AI inference, were available at only 4 of 10 sites. Participants prioritised AI for cancer screening, diagnosis, and biomarker prediction from H&E slides. Six implementation determinants were identified: AI literacy, governance frameworks, funding, evidence generation, clinical deployment, and collaboration. Financial constraints emerged as the dominant barrier across all determinants, and collaboration as a key enabler. Conclusion: AI implementation readiness varied substantially within income groups, suggesting that country income classification is an insufficient proxy for institutional readiness, underscoring the need for institution-level assessment. Realising the leapfrogging potential of AI in this context requires coordinated progress across all determinants, with sustainable financing, and collaboration as enablers of the other determinants. Key messages: What is already known on this topic?: Cancer incidence is rising rapidly in LMICs, where diagnostic capacity and molecular testing capabilities are severely constrained. AI could alleviate gaps, yet evidence on implementation determinants for integrating AI into cancer diagnostics in these settings remains limited. What this study adds?: In this exploratory sample, AI implementation readiness varied substantially across LMIC settings, including within the same country income group. Readiness is shaped by six implementation determinants: AI literacy, governance frameworks, funding, evidence generation, clinical deployment, and collaboration, with financial constraints as the dominant cross-cutting barrier. How this study might affect research, practice or policy?: These findings support institution-level feasibility assessment rather than reliance on country income classification as a proxy and highlight the need for sustainable financing and collaboration as foundational prerequisites for AI implementation in cancer diagnostics in LMICs.

Importance Advanced-stage disease at the time of diagnosis, with resultant high mortality, is among the most urgent issues for HIV-related Kaposi sarcoma (KS) in sub-Saharan Africa. Lack of access to skilled clinical personnel and histopathologic technology in the region contribute to diagnostic delays and advanced stage at diagnosis. Accordingly, new paradigms for KS diagnosis are needed. Objective To evaluate the accuracy of artificial intelligence (AI)-based interpretation of digital surface images of skin lesions to diagnose KS among dark-skinned patients in Uganda. Design Cross-sectional study of consecutive participants referred to skin biopsy services in Uganda because of clinical suspicion of KS. Lesions were photographed using a digital camera, and punch biopsies were obtained. Histopathologic interpretation was considered the gold standard. Using training (∼70% of images) and validation (∼10% of images) sets, we developed a prediction model using a rule-based combination of YOLO (You Only Look Once) version 5 and 8 object detection classifiers. Setting Free-of-charge skin biopsy services Participants Consecutive sample of 482 individuals were evaluated due to clinical suspicion of KS. Main Outcomes Sensitivity, specificity, positive and negative predictive value (with accompanying 95% confidence intervals) of the AI-based prediction model in a test set (∼20% of images). The accuracy of a dermatologist’s visual interpretation of images was also described. Results 472 participants (1385 images) were evaluable. Of these, 36% were female, median age was 34 years, and 94% had HIV; 332 had KS, and 140 had no KS by histopathology. In the test set, the AI-derived prediction model achieved 89% sensitivity (85%-94%) and 51% specificity (40%-61%) for diagnosing KS; positive predictive value was 81% (75%-86%) and negative predictive value was 67% (55%-78%). A dermatologist evaluating the same images, with emphasis on sensitivity, achieved sensitivity of 93% (89%-96%) and specificity of 19% (11%-28%). Conclusions and Relevance Among dark-skinned patients in Uganda with skin lesions suspicious for KS, evaluation of digital surface images by an AI-based prediction model produced moderate accuracy for diagnosing KS. While currently inadequate for clinical use, this inaugural assessment is sufficiently promising to justify evaluation of larger datasets and evolving technologies to determine if accuracy can be improved. Question Can an artificial intelligence (AI)-based prediction model be developed from digital images to accurately distinguish Kaposi sarcoma (KS) from non-KS in dark-skinned patients? Findings Evaluation of digital images of skin lesions from patients in Uganda by an AI-based prediction model produced moderate accuracy for diagnosing KS. Meaning In sub-Saharan Africa, where incidence and mortality from KS is high and delayed diagnosis is common due to limited specialized personnel and technical supplies, AI-based prediction models built on digital images taken of suspicious lesions may someday hasten KS diagnoses.

We examined the self-rated competence of Ugandan healthcare workers (HCWs) in obtaining informed consent for autopsies, considering the challenges of low autopsy acceptance rates globally. In September and October 2023, we conducted a nationwide cross-sectional study of HCWs, who provided informed consent to participate and completed an online, self-administered questionnaire. Participants’ self-rated competence in obtaining informed consent for autopsy was assessed through Likert scale questions. Knowledge and practices were also assessed. All scores were converted to percentages, with scores ≥80% indicating higher competence. We enrolled 216 HCWs (including 145 [67.1%] doctors), with a mean age of 31.6 ± 7.2 years. Overall, 55.6% (n = 120) had ever assisted in obtaining consent for autopsy, 43.6% (n = 100) had ever obtained consent for autopsy themselves, and 13.4% (n = 29) had ever attended training on obtaining consent for autopsy. The mean competency score was 59.8 ± 17.0% (perfect score, 100%), with 29 (13.4%) participants demonstrating high competence. Healthcare workers with adequate knowledge had higher competence scores (odds ratio [OR]: 15.0, 95% CI: 6.17–36.58, P P = 0.040). Fewer than one in five Ugandan HCWs demonstrated high self-rated competence or possessed adequate knowledge regarding informed consent for autopsies, and only a few had received specialized training on how to obtain consent for an autopsy. Therefore, there is a pressing need for enhanced training and increased awareness among Ugandan HCWs in obtaining informed consent for autopsies.

Background: BRCA1 gene dysfunction seen in epithelial ovarian carcinomas often results from germline mutations, somatic mutations, and promoter methylation. Identification of tumors with loss of BRCA1 protein expression has shown to have therapeutic and prognostic implications. The aim of this study was to determine the expression of BRCA1 protein in epithelial ovarian cancer (EOC) and the associated clinicopathological characteristics.Methods and Results: This was a cross‐sectional laboratory‐based study that used paraffin‐embedded tissue blocks of patients histologically diagnosed with EOC from January 2010 to August 2018. Tissue sections were stained with hematoxylin and eosin (H&E) for histological confirmation and with immunohistochemistry (IHC) using a mouse‐derived monoclonal antibody MS110 for BRCA1 protein expression. The association between BRCA1 protein expression and independent variables was determined using Pearson’s Chi‐square test. A total of 104 tissue blocks from patients with EOC were included in the study with a mean age of 48.7 ± 12.8 years. Serous tumors were the most common which comprised 74.0% (77/104) of all the tumors and majority of them 75.3% (58/77) were high grade. Loss of expression of BRCA1 protein expression was found in 33.7% (33/98) of all the cases. There was no statistically significant association between BRCA1 expression and age of patients, tumor grade, and histological subtype.Conclusion: There is a high expression of altered BRCA1 expression in tissues of EOC. Although it has not shown association with age of patients, histology types, and tumor grade, further studies need to assess its influence of the survival of cancer patients with EOC.

Background: Ultrasonography is a noninvasive modality for the initial assessment of thyroid nodules. The American College of Radiology Thyroid Imaging Reporting and Data System (ACR TI-RADS) has demonstrated good performance in differentiating malignant thyroid nodules. However, the combination of ACR TI-RADS categories and cytology has not been studied extensively, in Uganda. The study aims to correlate ACR TI-RADS with cytology among patients referred for US-guided fine-needle aspiration at Mulago National Referral Hospital. Methods: This was a hospital-based cross-sectional study that recruited 132 patients with thyroid nodules. Spearman’s correlation was used to establish a relationship between TI-RADS and cytology findings. The diagnostic accuracy of TI-RADS was assessed using sensitivity, specificity, positive and negative predictive values, and positive and negative likelihood ratios. Results: Of 132 study participants, 90% (n = 117) were females, and the mean age was 41 ± 13 years. One hundred sixty-one thyroid nodules were analyzed. More than half of the thyroid nodules (54.7%, n = 87) were solid or almost solid, 96.9% (n = 154) were shaped wider than tall, 57.2% (n = 91) had smooth margins, 83.7% (n = 133) were hyperechoic or isoechoic, and 88.7% (n = 141) had no echogenic foci. TI-RADS 3 was the most common at 42.9% (n = 69). The proportions of malignancy for TI-RADS 4 and TI-RADS 5 were 73.3% and 85.7%, respectively. The correlation between ACR TI-RADS and the Bethesda system of thyroid classification scores was r = 0.577. The sensitivity, specificity, positive and negative predictive values, and positive and negative likelihood ratios of ACR TI-RADS were 90.9%, 98.5%, 90%, 99.3%, 62.3, and 0.1, respectively. Conclusion: We found that ACR TI-RADS classification is an appropriate and noninvasive method for assessing thyroid nodules in routine practice. It can safely reduce the number of unnecessary fine-needle aspiration in a significant proportion of benign thyroid lesions. Thyroid nodules classified as TI-RADS 3 should be followed routinely. ACR TI-RADS should be standardized as the screening tool in resource-limited areas.

Kaposi’s sarcoma (KS) is an endothelial cancer caused by the Kaposi’s sarcoma-associated herpesvirus (KSHV) and is one of the most common cancers in sub-Saharan Africa. In limited-resource settings, traditional pathology infrastructure is often insufficient for timely diagnosis, leading to frequent diagnoses at advanced-stage disease where survival is poor. In this study, we investigate molecular diagnosis of KS performed in a point-of-care device to circumvent the limited infrastructure for traditional diagnosis. Using 506 mucocutaneous biopsies collected from patients at three HIV clinics in Uganda, we achieved 97% sensitivity, 92% specificity, and 96% accuracy compared to gold standard U.S.-based pathology. The results presented in this manuscript show that LAMP-based quantification of KSHV DNA extracted from KS-suspected biopsies has the potential to serve as a successful diagnostic for the disease and that diagnosis may be accurately achieved using a point-of-care device, reducing the barriers to obtaining KS diagnosis while increasing diagnostic accuracy.

Nakalanga syndrome is a clinical manifestation of onchocerciasis-associated epilepsy characterized by stunting, delayed or absent secondary sexual development and skeletal deformities, and is often accompanied by epileptic seizures. The pathophysiology of Nakalanga syndrome is unknown. Here, we describe the post-mortem findings of a 17-year-old female who died with Nakalanga syndrome in northern Uganda. Macroscopic and histopathological examination of all major organs (liver, lungs, kidney and heart), including the brain and the pituitary gland, was performed. The suspected cause of death was malaria, and all major organs and pituitary gland appeared normal, except the lungs, which were edematous consistent with the malaria. Neuropathological changes include signs of neuro-inflammation (gliosis and activated microglia), which co-localized with tau-reactive neurofibrillary tangles and threads. The pathology was most abundant in the frontal cortex, thalamic and hypothalamic regions, and mesencephalon. The choroid plexus showed psammoma bodies. These findings indicate accelerated aging, probably due to repeated seizures. The neuropathological findings were similar to other persons who died with onchocerciasis-associated epilepsy. Examination of the pituitary gland did not reveal new information concerning the underlying pathophysiological mechanism of Nakalanga syndrome. Therefore, more post-mortem studies should be performed.

The prevalence and distribution of histologically diagnosed breast disease are not well documented in low income countries, Uganda inclusive. Although the greater majority of breast lesions globally are benign, breast cancer is the most frequently diagnosed cancer all over the world. We aimed at documenting the prevalence of different breast diseases histologically diagnosed at the histopathology laboratory of the Department of Pathology of the Makerere University College of Health Sciences (MakCHS Lab) over a decade (2005–2014). We also describe the demographic characteristics of the patients in Uganda diagnosed with breast disease at the MakCHS Lab during the same period. This was a 10 year retrospective study of histologically diagnosed breast disease between 2005 and 2014 inclusive at the MakCHS Lab. We extracted information from hard copies of all 2510 histopathology reports retrieved from archives of the Department of Pathology at the MakCHS Lab. 640 records that were either damaged beyond recognition of key details, were duplicated, were implausible or had no conclusive diagnosis made were excluded. Information to be analyzed was then entered into Epidata (version 3.1) on a password protected laptop. Data analysis was done using SPSS software (v16 for Windows × 64). From the 1870 patients’ records eventually analyzed, breast disease was most diagnosed in female patients (97.1%). The overall mean age for breast disease diagnosis was 33 years (S.D ± 16.46) and median age 26 years (IQR: 20–43). Fibroadenoma (40.1%) was the most diagnosed breast disease overall. We noticed steadily increasing frequency of diagnosis of cancerous breast diseases over the last half of the study period. Invasive ductal carcinoma was the most diagnosed breast cancer (326 cases, 55.6%). A high female to male breast cancer ratio of 48:1 was observed. The highest regional breast cancer proportion was from the Western region of the Country. There is need for more research into the picture of breast disease in the country, covering various demographic characteristics of the country’s population for all regions and informing about its incidence rates and prevalence and also the breast cancer risk estimate for benign breast disease.

Percutaneous needle autopsy can overcome a number of barriers that limit the use of complete autopsies. We performed blind-and ultrasound guided needle autopsies in HIV-infected adults in Uganda. In this study we describe in detail the methods we used, the ability of both procedures to obtain sufficient tissue for further examination and the learning curve of the operators over time. If written informed consent was granted from the next of kin, we first performed a blind needle autopsy, puncturing brain, heart, lungs, liver, spleen and kidneys using predefined surface marking points. We then performed an ultrasound guided needle autopsy puncturing heart, liver, spleen and kidneys. The number of attempts, expected success and duration of the procedure were noted. A pathologist read the slides and indicated if the target tissue was present and of sufficient quality for pathological review. We report the predicted and true success rates, compare the yield of blind to ultrasound guided needle biopsies and evaluate the failure rate over time. Two operators performed 96 blind needle autopsies and 95 ultrasound guided needle autopsies. For blind needle biopsies true success rates varied from 56-99% and predicted success rates from 89-99%. For ultrasound guided needle biopsies true success rates varied from 72-100% and predicted success rates from 84-98%. Ultrasound guidance led to a significantly higher success rate in heart and left kidney. A learning curve was observed over time with decreasing failure rates with increasing experience and a shorter duration of the needle autopsy. Needle autopsy can successfully obtain tissue for further pathological review in the vast majority of cases, with a decrease in failure rate with increasing experience of the operator. The benefit of ultrasound guidance will depend on the population, the disease and organ of interest and the local circumstances. Our results justify further evaluation of needle autopsies as a method to establish a cause of death.