Nakhleh, Raouf E.

Abstract Aims Current non-invasive screening methods for cardiac allograft rejection have shown limited discrimination and are yet to be broadly integrated into heart transplant care. Given electrocardiogram (ECG) changes have been reported with severe cardiac allograft rejection, this study aimed to develop a deep-learning model, a form of artificial intelligence, to detect allograft rejection using the 12-lead ECG (AI-ECG). Methods and results Heart transplant recipients were identified across three Mayo Clinic sites between 1998 and 2021. Twelve-lead digital ECG data and endomyocardial biopsy results were extracted from medical records. Allograft rejection was defined as moderate or severe acute cellular rejection (ACR) based on International Society for Heart and Lung Transplantation guidelines. The extracted data (7590 unique ECG-biopsy pairs, belonging to 1427 patients) was partitioned into training (80%), validation (10%), and test sets (10%) such that each patient was included in only one partition. Model performance metrics were based on the test set (n = 140 patients; 758 ECG-biopsy pairs). The AI-ECG detected ACR with an area under the receiver operating curve (AUC) of 0.84 [95% confidence interval (CI): 0.78–0.90] and 95% (19/20; 95% CI: 75–100%) sensitivity. A prospective proof-of-concept screening study (n = 56; 97 ECG-biopsy pairs) showed the AI-ECG detected ACR with AUC = 0.78 (95% CI: 0.61–0.96) and 100% (2/2; 95% CI: 16–100%) sensitivity. Conclusion An AI-ECG model is effective for detection of moderate-to-severe ACR in heart transplant recipients. Our findings could improve transplant care by providing a rapid, non-invasive, and potentially remote screening option for cardiac allograft function.

Abstract Background Current approaches utilizing non-invasive methods to screen for cardiac allograft rejection (gene expression profiling and cell free DNA) have yet to be broadly integrated into heart transplant management and have shown limited discrimination (AUCs of 0.68 and 0.77, respectively). Changes in the electrocardiogram (ECG) have been reported at the time of severe cardiac rejection, including low voltages and conduction abnormalities. It remains unknown if subtle ECG changes correlating with cardiac allograft rejection can be detected earlier using machine learning methods. Purpose We sought to develop an artificial intelligence (AI) model to detect cardiac allograft rejection based on the 12 lead ECG. Methods We identified all patients who underwent a heart transplant at 3 hospital sites within a single health system from Jan 1998 through Apr 2021 and extracted digital 12-lead ECG data as well as endomyocardial biopsy pathology results from the electronic medical record. We partitioned our data into a training (80%), validation (10%), and test set (10%) based on a group stratification sampling. Each patient was present in only one set and each set had a positivity rate of 2.6% with 6,074/758/758 ECGs belonging to 1,146/140/141 unique patients in each set respectively. Cardiac allograft rejection was defined as moderate or severe acute cellular rejection based on International Society for Heart and Lung Transplantation (ISHLT) guidelines. A convolutional neural network, using the 12-lead ECG data as input, was trained with hyperparameter optimization for regularization, learning rate adjustments, and class weights. Model performance metrics were based on the test data and estimated using the final model architecture. Results 1,587 heart transplant recipients who had at least one endomyocardial biopsy were evaluated for inclusion. We limited our sample to ECGs performed within 30 days of the biopsy date (7,590 ECGs, representing 1,425 unique patients). Our study population had a median age of 55.8 years and 28.7% were female. The median number of ECG-biopsy pairs per patient was 5. The majority of endomyocardial biopsy results were classified as none or mild rejection (97.1%), and 2.9% had moderate/severe rejection. The ECG-based AI model detected cardiac allograft rejection with an area under the receiver operative curve (AUC) of 0.84 in the test set. The sensitivity, specificity, positive and negative predictive values were 95%, 52.6%. 5.2% and 99.7% respectively. Conclusions An AI-ECG model appears to outperform novel non-invasive laboratory tests (gene expression profiling and cell free DNA) for detecting cardiac allograft rejection and does not require a blood draw or the additional complexities surrounding sample processing. This model relies on a readily available and relatively inexpensive test, the ECG. In addition, AI predictions can be made available within a few minutes following ECG acquisition. Funding Acknowledgement Type of funding sources: Private hospital(s). Main funding source(s): Mayo Clinic

ABSTRACT Novel chemotherapeutic agents are developed to treat recurrent/relapsed lymphoid malignancies. Umbralisib, a novel phosphatidylinositol 3-kinase inhibitor with a selective isoform binding, has shown an improved efficacy and safety profile in clinical trials. Immune-mediated colitis, a frequently observed dose-limiting adverse event of phosphatidylinositol 3-kinase inhibitors, has been mostly observed at supratherapeutic doses in the trials, with grade 1 or 2 diarrhea being the most common adverse event at the therapeutic dose (800 mg PO QD). We present a grade-3 colitis that can be attributed to umbralisib-mediated immune toxicity in a patient with chronic lymphocytic leukemia at the therapeutic dose.

Abstract Context. —The Joint Commission and the College of American Pathologists have emphasized improved communication as a strategy to improve patient safety and reduce errors. Objective. —To determine current policies and practices concerning critical and/or significant and unexpected results in anatomic pathology. Design. —A survey was distributed with the 2007 D mailing of the Performance Improvement Program slides. The survey included questions that determined laboratory size, practice setting, and anatomic pathology critical and/or significant and unexpected result policies and practices. Results. —Surveys from 1130 laboratories were received. A total of 75% had a written policy regarding anatomic pathology critical and/or significant and unexpected results; 25% did not. A total of 30% of laboratories with written policies stated that their policies included guidelines but did not include specific examples. A total of 33% listed 5 or fewer specific examples, 18% listed more than 5 examples, and 19% stated that they had a specifically defined list of significant and unexpected and/or critical diagnoses. The conditions that were listed included malignancies (48% of all laboratories), findings not expected by the clinical history (45%), life-threatening infections (45%), no chorionic villi in products of conception (37%), inflammatory or immunologic processes (19%), and organ rejection (14%). Laboratories with a higher median number of accessioned surgical and cytology cases and independent laboratories tended to have policies with more than 5 specific examples or precise lists of must-call diagnoses ( P < .001). Conclusions. —This survey illustrates current anatomic pathology policies and practices with respect to critical and significant and unexpected results.

Abstract Context. —Monitoring customer satisfaction is a valuable component of a laboratory quality improvement program. Objective. —To survey the level of physician satisfaction with hospital clinical laboratory services. Design. —Participating institutions provided demographic and practice information and survey results of physician satisfaction with defined aspects of clinical laboratory services, rated on a scale of 1 (poor) to 5 (excellent). Results. —One hundred thirty-eight institutions participated in this study and submitted a total of 4329 physician surveys. The overall satisfaction score for all institutions ranged from 2.9 to 5.0. The median overall score for all participants was 4.1 (10th percentile, 3.6; 90th percentile, 4.5). Physicians were most satisfied with the quality/reliability of results and staff courtesy, with median values of excellent or good ratings of 89.9%. Of the 5 service categories that received the lowest percentage values of excellent/good ratings (combined scores of 4 and 5), 4 were related to turnaround time for inpatient stat, outpatient stat, routine, and esoteric tests. Surveys from half of the participating laboratories reported that 96% to 100% of physicians would recommend the laboratory to other physicians. The category most frequently selected as the most important category of laboratory services was quality/reliability of results (31.7%). Conclusions. —There continues to be a high level of physician satisfaction and loyalty with clinical laboratory services. Test turnaround times are persistent categories of dissatisfaction and present opportunities for improvement.

Abstract Context.—National patient safety goals and error reduction efforts should be addressed by each surgical pathology laboratory. Objective.—To review issues relevant to patient safety and error reduction in surgical pathology in the context of continuous quality improvement. Data Sources.—The literature is reviewed. Conclusions.—Patient safety goals can and should be addressed within the context of a quality improvement plan. Multiple factors that contribute to errors in surgical pathology are discussed. The current literature defines the extent of these problems within specific segments of the test cycle (preanalytic, analytic, and postanalytic). Potential solutions are presented that may reduce or avoid errors. In addition, general principles are outlined that enhance the laboratory's ability to successfully and continuously address patient safety and error reduction.

Abstract Context. —There are multiple elements that can be measured to assess the quality of a surgical pathology laboratory. Overall customer satisfaction is an excellent “global” measure, because it highlights the unique insight of laboratory performance from the users' perspective. Objective. —To measure customer satisfaction with surgical pathology reports. Design. —This study was based on a subscription Q-Tracks study. Voluntary participants were asked to distribute and collect a minimum of 25 surveys per quarter from their clients. Four parameters were graded, which included overall satisfaction, report turnaround time (TAT), completeness, and style on a scale of 1 (poor) to 5 (excellent). Each laboratory submitted quarterly data to the College of American Pathologists, where the data were tabulated and analyzed. Each laboratory could compare their performance in all 4 measures against the entire cohort or a selected subgroup of laboratories. Overall customer satisfaction with surgical pathology reports and 3 subcategories of report TAT, completeness, and style were the main outcome measures. Results. —This study ran during 2004 and 2005, with 41 and 33 participant laboratories, respectively. The median score for overall satisfaction, TAT, completeness, and style were 4.57, 4.31, 4.62, and 4.64 in 2004, and 4.64, 4.56, 4.65, and 4.68 in 2005, respectively. Most laboratories reported results for 4 quarters or fewer. There was no statistically significant change in overall satisfaction over time. Conclusions. —Overall satisfaction scores for surgical pathology reports as well as satisfaction with report TAT, completeness, and style were high. Report TAT received the lowest scores of all parameters.

Abstract Context. —Because of its complex nature, surgical pathology practice is inherently error prone. Currently, there is pressure to reduce errors in medicine, including pathology. Objective. —To review factors that contribute to errors and to discuss error-reduction strategies. Design. —Literature review. Results. —Multiple factors contribute to errors in medicine, including variable input, complexity, inconsistency, tight coupling, human intervention, time constraints, and a hierarchical culture. Strategies that may reduce errors include reducing reliance on memory, improving information access, error-proofing processes, decreasing reliance on vigilance, standardizing tasks and language, reducing the number of handoffs, simplifying processes, adjusting work schedules and environment, providing adequate training, and placing the correct people in the correct jobs. Conclusions. —Surgical pathology is a complex system with ample opportunity for error. Significant error reduction is unlikely to occur without a sustained comprehensive program of quality control and quality assurance. Incremental adoption of information technology and automation along with improved training in patient safety and quality management can help reduce errors.

Abstract Context.—Anatomic pathology discrepancy frequencies have not been rigorously studied. Objective.—To determine the frequency of anatomic pathology discrepancies and the causes of these discrepancies. Design.—Participants in the College of American Pathologists Q-Probes program self-reported the number of anatomic pathology discrepancies in their laboratories by prospectively performing secondary review (post–sign-out) of 100 surgical pathology or cytology specimens. Reasons for the secondary review included conferences, external review, internal quality assurance policy, and physician request. Participants.—Seventy-four laboratories self-reported data. Main Outcome Measures.—Frequency of anatomic pathology discrepancy; type of discrepancy (ie, change in margin status, change in diagnosis, change in patient information, or typographic error); effect of discrepancy on patient outcome (ie, no harm, near miss, or harm); and clarity of report. Results.—The mean and median laboratory discrepancy frequencies were 6.7% and 5.1%, respectively. Forty-eight percent of all discrepancies were due to a change within the same category of interpretation (eg, 1 tumor type was changed to another tumor type). Twenty-one percent of all discrepancies were due to a change across categories of interpretation (eg, a benign diagnosis was changed to a malignant diagnosis). Although the majority of discrepancies had no effect on patient care, 5.3% had a moderate or marked effect on patient care. Conclusions.—This study establishes a mean multi-institutional discrepancy frequency (related to secondary review) of 6.7%.