Citations

Abstract Innovation in medical imaging artificial intelligence (AI)/machine learning (ML) demands extensive data collection, algorithmic advancements, and rigorous performance assessments encompassing aspects such as generalizability, uncertainty, bias, fairness, trustworthiness, and interpretability. Achieving widespread integration of AI/ML algorithms into diverse clinical tasks will demand a steadfast commitment to overcoming issues in model design, development, and performance assessment. The complexities of AI/ML clinical translation present substantial challenges, requiring engagement with relevant stakeholders, assessment of cost-effectiveness for user and patient benefit, timely dissemination of information relevant to robust functioning throughout the AI/ML lifecycle, consideration of regulatory compliance, and feedback loops for real-world performance evidence. This commentary addresses several hurdles for the development and adoption of AI/ML technologies in medical imaging. Comprehensive attention to these underlying and often subtle factors is critical not only for tackling the challenges but also for exploring novel opportunities for the advancement of AI in radiology.

Abstract Background: Hepatocellular carcinoma (HCC) is one of the leading causes of cancer-related mortality worldwide. In sub-Saharan Africa, HCC is the second most common cancer in men and the sixth in women. African patients present with HCC at a young age and at advanced stage of disease. Improved surveillance can increase early HCC detection. The GALAD score has good sensitivity in the diagnosis of HCC in Asia, Europe and North America, however, it has not been validated in an African cohort. The aim of this study was to assess the performance of the GALAD score in the diagnosis of HCC in Ghanaian patients. Methods: Clinical data from patients with cirrhosis (n=93) or HCC (n=78) from out-patient hepatology clinics at 3 teaching hospitals in Ghana were abstracted, and serum samples analyzed. A logistic regression model predicting HCC status based on GALAD score was constructed to obtain the Receiver Operating Characteristics (ROC) curve for GALAD. The area under the curve (AUC) with 95% confidence interval was calculated and used to assess the performance of the GALAD score in surveillance for HCC. Results: The median AFP (5010.2 vs. 3.4), L3% (28.3 vs. 0.3) and DCP (507.6 vs. 0.2) levels were all significantly higher among HCC cases than the controls (p<0.0001). The median GALAD score was also higher among HCC patients (8.0 vs. -4.1, p<0.0001). The AUC of the GALAD score in the diagnosis of HCC was 0.86 (95% CI 0.79 - 0.92). At a cut-off value of -0.37, the GALAD score had a sensitivity of 0.81 and a specificity of 0.86. In a re-fit of the logistic regression model, the most significant factors in the GALAD model in our cohort were AFP-L3% and log(DCP). The AUC of the new GALAD model after model selection was 0.89. Conclusion: The GALAD score can be used to detect HCC in Ghana and has the potential to improve HCC surveillance in low-resource settings in sub-Saharan Africa. Citation Format: Yvonne Ayerki Nartey, Ju Dong Yang, Tyler Zemla, Joshua Ayawin, Shadrack Osei Asibey, Mohamed El-Kassas, Sally Afua Bampoh, Amoako Duah, Adwoa Agyei-Nkansah, Yaw Asante Awuku, Mary Yeboah Afihene, Hiroyuki Yamada, Jun Yin, Amelie Plymoth, Lewis Rowland Roberts. Accuracy of the GALAD serologic model in the diagnosis of hepatocellular carcinoma in Ghanaian patients [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 4802.

Abstract A neurological dogma is that the contralateral effects of brain injury are set through crossed descending neural tracts. We have recently identified a novel topographic neuroendocrine system (T-NES) that operates via a humoral pathway and mediates the left-right side-specific effects of unilateral brain lesions. In rats with completely transected thoracic spinal cords, unilateral injury to the sensorimotor cortex produced contralateral hindlimb flexion, a proxy for neurological deficit. Here, we investigated in acute experiments whether T-NES consists of left and right counterparts and whether they differ in neural and molecular mechanisms. We demonstrated that left- and right-sided hormonal signaling is differentially blocked by the δ-, κ- and µ-opioid antagonists. Left and right neurohormonal signaling differed in targeting the afferent spinal mechanisms. Bilateral deafferentation of the lumbar spinal cord abolished the hormone-mediated effects of the left-brain injury but not the right-sided lesion. The sympathetic nervous system was ruled out as a brain-to-spinal cord-signaling pathway since hindlimb responses were induced in rats with cervical spinal cord transections that were rostral to the preganglionic sympathetic neurons. Analysis of gene–gene co-expression patterns identified the left- and right-side-specific gene co-expression networks that were coordinated via the humoral pathway across the hypothalamus and lumbar spinal cord. The coordination was ipsilateral and disrupted by brain injury. These findings suggest that T-NES is bipartite and that its left and right counterparts contribute to contralateral neurological deficits through distinct neural mechanisms, and may enable ipsilateral regulation of molecular and neural processes across distant neural areas along the neuraxis.

Abstract The adoption of artificial intelligence (AI) tools in medicine poses challenges to existing clinical workflows. This commentary discusses the necessity of context-specific quality assurance (QA), emphasizing the need for robust QA measures with quality control (QC) procedures that encompass (1) acceptance testing (AT) before clinical use, (2) continuous QC monitoring, and (3) adequate user training. The discussion also covers essential components of AT and QA, illustrated with real-world examples. We also highlight what we see as the shared responsibility of manufacturers or vendors, regulators, healthcare systems, medical physicists, and clinicians to enact appropriate testing and oversight to ensure a safe and equitable transformation of medicine through AI.

Abstract To elucidate the molecular basis of multiple system atrophy (MSA), a neurodegenerative disease, we conducted a genome-wide association study (GWAS) in a Japanese MSA case/control series followed by replication studies in Japanese, Korean, Chinese, European, and North American samples. In the GWAS stage, rs2303744 on chromosome 19 showed a suggestive association (P = 6.5 × 10− 7) that was replicated in additional Japanese samples (P = 2.9 × 10− 6, odds ratio (OR) = 1.58; 95% confidence interval (CI), 1.30–1.91), and then confirmed as highly significant in a meta-analysis of East Asian population data (P = 5.0 × 10− 15, OR = 1.49; 95% CI 1.35–1.65). The association of rs2303744 with MSA remained significant in combined European/North American samples (P = 0.017, OR = 1.12; 95% CI 1.02–1.23) despite allele frequencies being quite different between these populations. rs2303744 leads to an amino acid substitution in PLA2G4C that encodes the cPLA2γ lysophospholipase/transacylase. The cPLA2γ-Ile143 isoform encoded by the MSA risk allele has significantly decreased transacylase activity compared with the alternate cPLA2γ-Val143 isoform, and may perturb membrane phospholipids and promote oligodendroglial pathology.

Background The C-C motif chemokine receptor 5 (CCR5)/C-C motif chemokine ligand 5 (CCL5) axis plays a major role in colorectal cancer (CRC). We aimed to characterize the molecular features associated with CCR5 / CCL5 expression in CRC and to determine whether CCR5 / CCL5 levels could impact treatment outcomes. Methods 7604 CRCs tested with NextGen Sequencing on DNA and RNA were analyzed. Molecular features were evaluated according to CCR5 and CCL5 tumor gene expression quartiles. The impact on treatment outcomes was assessed in two cohorts, including 6341 real-world patients and 429 patients from the Cancer and Leukemia Group B (CALGB)/SWOG 80405 trial. Results CCR5 / CCL5 expression was higher in right-sided versus left-sided tumors, and positively associated with consensus molecular subtypes 1 and 4. Higher CCR5 / CCL5 expression was associated with higher tumor mutational burden, deficiency in mismatch repair and programmed cell death ligand 1 (PD-L1) levels. Additionally, high CCR5 / CCL5 were associated with higher immune cell infiltration in the tumor microenvironment (TME) of MMR proficient tumors. Ingenuity pathway analysis revealed upregulation of the programmed cell death protein 1 (PD-1)/PD-L1 cancer immunotherapy pathway, phosphatase and tensin homolog (PTEN) and peroxisome proliferator-activated receptors (PPAR) signaling, and cytotoxic T-lymphocyte antigen 4 (CTLA-4) signaling in cytotoxic T lymphocytes, whereas several inflammation-related pathways were downregulated. Low CCR5 / CCL5 expression was associated with increased benefit from cetuximab-FOLFOX treatment in the CALGB/SWOG 80405 trial, where significant treatment interaction was observed with biologic agents and chemotherapy backbone. Conclusions Our data show a strong association between CCR5/CCL5 gene expression and distinct molecular features, gene expression profiles, TME cell infiltration, and treatment benefit in CRC. Targeting the CCR5/CCL5 axis may have clinical applications in selected CRC subgroups and may play a key role in developing and deploying strategies to modulate the immune TME for CRC treatment.

Introduction: The impact of COVID-19 was felt across cardiology as resources were redirected to critical care, delaying elective cardiology procedures. Hypothesis: To evaluate differences in outcomes of cardiac implantable electronic device (CIED) procedures in 2020 (height of COVID) vs. 2016-2019 (pre-COVID). Methods: Adults >= 18 years undergoing CIED procedures were identified from National Inpatient Sample Databases (NIS) 2016-2020. Complications and mortality were assessed based on the type of device: Permanent Pacemaker (PPM); Implantable Cardioverter Defibrillator (ICD); Cardiac Resynchronization Therapy with Pacemaker (CRT-P) or defibrillator (CRT-D): Leadless PPM; Generator explant (GE), and Lead extraction (LE). Results: A 13.8% drop in CIED admissions was recorded between 2019 and 2020 (p<0.01). Fewer procedures involving PPM, ICD, CRT-P, CRT-D, LE, and GE were observed between 2019 and 2020. Leadless PPM utilization rose consistently between 2016-2020 (47.6% increase during 2020 vs 2019)(Figure). In 2020, a higher proportion of patients were classified with a Charlson Comorbidity Index score of ≥3 (severe) (47.4% in 2020 vs. 42.9% in 2016-2019, p<0.01). The table shows significantly higher complications during CIED implants with no changes in mortality related to procedures. Conclusions: Patients admitted for CIED implants in 2020 had more severe comorbidities and expressed considerable differences in complications than pre-COVID era (2016-2019), but mortality remained unchanged. Leadless PPM use increased during pandemic, as against a decrease in the use of other CIEDs.

Introduction: While Coronavirus disease 2019 (COVID-19) primarily affects the respiratory system but can impact the cardiovascular system can triggering various complications. Hypothesis: To investigate the trend in age-adjusted mortality rates (AAMRs) in patients with atrial fibrillation(AF) among COVID-19-positive patients (CPP) by race and sex. Methods: Trends in AAMRs (per 100,000) from all-cause in patients with AF between April to December 2016-2020 were evaluated via the National Inpatient Sample and the US Census Bureau. The AAMRs among CPP with AF in 2020 was also evaluated between racial and sex cohorts. Results: A gradual rise in AAMR(all causes) in patients with AF was observed between 2016-2019, with a 49.4% overall rise between 2019(271 per 100,000) and 2020(405 per 100,000) ( Table 1a and Figure ). Around 250475 COVID-19-positive patients had AF, and 24.2% of cases did not survive their hospitalization (vs. 10.9% in those without AF, p<0.01, aOR 1.457, 95% CI 1.438-1.477). Among CPP males and females with AF that died, White females and males were the oldest, while Black males and females reported the highest mean Charlson Comorbidity Index (CCI) score ( Table 1b ).Furthermore, among the CPP with AF in 2020, Hispanics reported a higher AAMR (119.6 per 100,000) than Whites(111.1 per 100,000) and Blacks(101.9 per 100,000). AAMRs in males were higher in Hispanic males (144.8 per 100,000), followed by White males (129.7 per 100,000), and Black males (110 per 100,000). However, the AAMRs in females were highest in Black females (91.3 per 100,000), followed by White females (88.3 per 100,000) and Hispanic females (87.0 per 100,000). ( Table 1c ) Conclusions: AAMRs in patients with AF are on the rise and a sharp growth was seen during the first year of the pandemic. AAMRs among CPP patients with AF were highest among Black females and Hispanic males, while White males and females were oldest and Black males and females reported the highest CCI scores.