RESUMEN
BACKGROUND: During a pandemic, it is important for clinicians to stratify patients and decide who receives limited medical resources. Machine learning models have been proposed to accurately predict COVID-19 disease severity. Previous studies have typically tested only one machine learning algorithm and limited performance evaluation to area under the curve analysis. To obtain the best results possible, it may be important to test different machine learning algorithms to find the best prediction model. OBJECTIVE: In this study, we aimed to use automated machine learning (autoML) to train various machine learning algorithms. We selected the model that best predicted patients' chances of surviving a SARS-CoV-2 infection. In addition, we identified which variables (ie, vital signs, biomarkers, comorbidities, etc) were the most influential in generating an accurate model. METHODS: Data were retrospectively collected from all patients who tested positive for COVID-19 at our institution between March 1 and July 3, 2020. We collected 48 variables from each patient within 36 hours before or after the index time (ie, real-time polymerase chain reaction positivity). Patients were followed for 30 days or until death. Patients' data were used to build 20 machine learning models with various algorithms via autoML. The performance of machine learning models was measured by analyzing the area under the precision-recall curve (AUPCR). Subsequently, we established model interpretability via Shapley additive explanation and partial dependence plots to identify and rank variables that drove model predictions. Afterward, we conducted dimensionality reduction to extract the 10 most influential variables. AutoML models were retrained by only using these 10 variables, and the output models were evaluated against the model that used 48 variables. RESULTS: Data from 4313 patients were used to develop the models. The best model that was generated by using autoML and 48 variables was the stacked ensemble model (AUPRC=0.807). The two best independent models were the gradient boost machine and extreme gradient boost models, which had an AUPRC of 0.803 and 0.793, respectively. The deep learning model (AUPRC=0.73) was substantially inferior to the other models. The 10 most influential variables for generating high-performing models were systolic and diastolic blood pressure, age, pulse oximetry level, blood urea nitrogen level, lactate dehydrogenase level, D-dimer level, troponin level, respiratory rate, and Charlson comorbidity score. After the autoML models were retrained with these 10 variables, the stacked ensemble model still had the best performance (AUPRC=0.791). CONCLUSIONS: We used autoML to develop high-performing models that predicted the survival of patients with COVID-19. In addition, we identified important variables that correlated with mortality. This is proof of concept that autoML is an efficient, effective, and informative method for generating machine learning-based clinical decision support tools.
Asunto(s)
COVID-19/mortalidad , Aprendizaje Automático , COVID-19/virología , Femenino , Humanos , Masculino , Persona de Mediana Edad , Modelos Estadísticos , Pandemias , Estudios Retrospectivos , SARS-CoV-2/aislamiento & purificación , Análisis de SupervivenciaRESUMEN
OBJECTIVE: Persistent infection with oncogenic high risk HPV (hrHPV) types causes virtually all cases of cervical cancer. HPV 16 and 18 have been targeted for individual genotyping and vaccination because of their presence in 71% of invasive cervical cancers worldwide. Montefiore Medical Center, Bronx, New York serves a population known for ethnic and racial diversity. Given this diversity it is possible that HPV genotypes not individually detected by current testing are causing significant disease. METHODS: We conducted a retrospective analysis of liquid based cervicovaginal cytology and Cobas HPV results reported between October 5, 2015 and March 30, 2016. This included 20 483 samples from patients aged 16-95 (average age 42), with racial distribution including: African-American 32.4%, Other (includes denied, unknown, mixed, Hispanic) 52.1%, Caucasian 14.5%, Asian 0.7%, American Indian/Alaskan Native 0.3%. In all, 14 938 samples (72.9%) were submitted for clinically requested COBAS 4800 HPV testing, which separately reports HPV 16, 18 and a pool of 12 other hrHPV. RESULTS: A total of 3180 (21.5%) tested hrHPV positive. The percentage of patients with cytologic diagnosis of HSIL (high-grade squamous intraepithelial lesion) that were positive only for HPV 16 was 19.4% vs 1.8% for all cytologic diagnoses. However, only one of the HSIL cases was HPV 18 positive along with other hrHPV (OHR). Surprisingly, a majority (64.5%) was positive for only OHR. CONCLUSIONS: Further evaluation is needed to determine if this pool of other hrHPV includes individual genotypes that in our population carry a higher risk of persistence and progression to cancer.
Asunto(s)
Detección Precoz del Cáncer , Lesiones Intraepiteliales Escamosas de Cuello Uterino/diagnóstico , Displasia del Cuello del Útero/diagnóstico , Neoplasias del Cuello Uterino/diagnóstico , Adolescente , Adulto , Anciano , Anciano de 80 o más Años , Femenino , Genotipo , Papillomavirus Humano 16/genética , Papillomavirus Humano 16/aislamiento & purificación , Papillomavirus Humano 18/genética , Papillomavirus Humano 18/aislamiento & purificación , Humanos , Tamizaje Masivo , Persona de Mediana Edad , Ciudad de Nueva York/epidemiología , Prueba de Papanicolaou , Infecciones por Papillomavirus , Lesiones Intraepiteliales Escamosas de Cuello Uterino/epidemiología , Lesiones Intraepiteliales Escamosas de Cuello Uterino/genética , Lesiones Intraepiteliales Escamosas de Cuello Uterino/virología , Neoplasias del Cuello Uterino/epidemiología , Neoplasias del Cuello Uterino/genética , Neoplasias del Cuello Uterino/virología , Adulto JovenRESUMEN
With the global outbreak of COVID-19, the demand for testing rapidly increased and quickly exceeded the testing capacities of many laboratories. Clinical tests which receive CE (Conformité Européenne) and Food and Drug Administration (FDA) authorisations cannot always be tested thoroughly in a real-world environment. Here we demonstrate the long-term stability of nasopharyngeal swab specimens for SARS-CoV-2 molecular testing across three assays recently approved by the US FDA under Emergency Use Authorization. This study demonstrates that nasopharyngeal swab specimens can be stored under refrigeration or even ambient conditions for 21 days without clinically impacting the results of the real-time reverse transcriptase-PCR testing.
Asunto(s)
COVID-19/diagnóstico , SARS-CoV-2/aislamiento & purificación , Manejo de Especímenes/métodos , COVID-19/virología , Prueba de Ácido Nucleico para COVID-19 , Humanos , Laboratorios de Hospital , Nasofaringe/virología , Refrigeración , SARS-CoV-2/genética , Factores de TiempoRESUMEN
OBJECTIVE: To compare the cytologic preparations of 130 cervical specimens (from women of various ethnicities at high risk for human papillomavirus [HPV] infection) using the SurePath (SP) collection system with specimens gathered using the ThinPrep (TP) system, as processed on the Cobas 4800 analyzer, to determine which collection method more accurately identifies HPV infection. METHODS: In our prospective study, specimens were collected from 130 women of various ethnicities residing in or near Bronx County, NY. The SP-collected specimen was first processed for cytologic findings; if clinical HPV testing was requested on that specimen, it was tested using Hybrid Capture II (HC2) methodology. We tested the remnant SP-collected cell concentrate using the Cobas analyzer. Then, the TP-collected and SP-collected specimens were tested in the same run on that analyzer, and the results were compared. We also compared the results with the concurrent cytologic findings. RESULTS: The results were concordant for overall HR-HPV status in 93.8% of cases. Also, a statistically significant lower cycle threshold value was observed with Cobas testing of specimen concentrates tested via the BD SurePath Pap Test (P = .001), suggesting higher sensitivity compared with specimens tested via the ThinPrep Pap Test. CONCLUSION: Cobas 4800 HPV testing of SP-collected specimen concentrates yields comparable results to TP-collected specimen concentrates. Based on the limited data that we derived, SP collection may be a more favorable methodology than TP collection for HPV testing of individuals at high risk in our ethnically diverse, urban patient population.