Artificial intelligence-based models enabling accurate diagnosis of ovarian cancer using laboratory tests in China: a multicentre, retrospective cohort study.

BACKGROUND: Ovarian cancer is the most lethal gynecological malignancy. Timely diagnosis of ovarian cancer is difficult due to the lack of effective biomarkers. Laboratory tests are widely applied in clinical practice, and some have shown diagnostic and prognostic relevance to ovarian cancer. We aimed to systematically evaluate the value of routine laboratory tests on the prediction of ovarian cancer, and develop a robust and generalisable ensemble artificial intelligence (AI) model to assist in identifying patients with ovarian cancer. METHODS: In this multicentre, retrospective cohort study, we collected 98 laboratory tests and clinical features of women with or without ovarian cancer admitted to three hospitals in China during Jan 1, 2012 and April 4, 2021. A multi-criteria decision making-based classification fusion (MCF) risk prediction framework was used to make a model that combined estimations from 20 AI classification models to reach an integrated prediction tool developed for ovarian cancer diagnosis. It was evaluated on an internal validation set (3007 individuals) and two external validation sets (5641 and 2344 individuals). The primary outcome was the prediction accuracy of the model in identifying ovarian cancer. FINDINGS: Based on 52 features (51 laboratory tests and age), the MCF achieved an area under the receiver-operating characteristic curve (AUC) of 0·949 (95% CI 0·948-0·950) in the internal validation set, and AUCs of 0·882 (0·880-0·885) and 0·884 (0·882-0·887) in the two external validation sets. The model showed higher AUC and sensitivity compared with CA125 and HE4 in identifying ovarian cancer, especially in patients with early-stage ovarian cancer. The MCF also yielded acceptable prediction accuracy with the exclusion of highly ranked laboratory tests that indicate ovarian cancer, such as CA125 and other tumour markers, and outperformed state-of-the-art models in ovarian cancer prediction. The MCF was wrapped as an ovarian cancer prediction tool, and made publicly available to provide estimated probability of ovarian cancer with input laboratory test values. INTERPRETATION: The MCF model consistently achieved satisfactory performance in ovarian cancer prediction when using laboratory tests from the three validation sets. This model offers a low-cost, easily accessible, and accurate diagnostic tool for ovarian cancer. The included laboratory tests, not only CA125 which was the highest ranked laboratory test in importance of diagnostic assistance, contributed to the characterisation of patients with ovarian cancer. FUNDING: Ministry of Science and Technology of China; National Natural Science Foundation of China; Natural Science Foundation of Guangdong Province, China; and Science and Technology Project of Guangzhou, China. TRANSLATION: For the Chinese translation of the abstract see Supplementary Materials section.

Timing of interval debulking surgery and postoperative chemotherapy after neoadjuvant chemotherapy in advanced epithelial ovarian cancer: a multicenter real-world study.

BACKGROUND: To investigate the prognostic relevance of the time to interval debulking surgery (TTS) and the time to postoperative adjuvant chemotherapy (TTC) after the completion of neoadjuvant chemotherapy (NACT). METHODS: A retrospective real-word study included 658 patients with histologically confirmed advanced epithelial ovarian cancer who received NACT at seven tertiary hospitals in China from June 2008 to June 2020. TTS was defined as the time interval from the completion of NACT to the time of interval debulking surgery (IDS). TTC was defined as the time interval from the completion of NACT to the initiation of postoperative adjuvant chemotherapy (PACT). RESULTS: The median TTS and TTC were 25 (IQR, 20-29) and 40 (IQR, 33-49) days, respectively. Patients with TTS > 25 days were older (55 vs. 53 years, P = 0.012) and received more NACT cycles (median, 3 vs. 2, P = 0.002). Similar results were observed in patients with TTC > 40 days. In the multivariate analyses, TTS and TTC were not associated with PFS when stratified by median, quartile, or integrated as continuous variables (all P > 0.05). However, TTS and TTC were significantly associated with worse OS when stratified by median (P = 0.018 and 0.018, respectively), quartile (P = 0.169, 0.014, 0.027 and 0.012, 0.001, 0.033, respectively), or integrated as continuous variables (P = 0.018 and 0.011, respectively). Similarly, increasing TTS and TTC intervals were associated with a higher risk of death (Ptrend = 0.016 and 0.031, respectively) but not with recurrence (Ptrend = 0.103 and 0.381, respectively). CONCLUSION: The delays of IDS and PACT after the completion of NACT have adverse impacts on OS but no impacts on PFS, which indicates that reducing delays of IDS and PACT might ameliorate the outcomes of ovarian cancer patients treated with NACT.

Development and validation of a nomogram to predict cancer-specific survival of mucinous epithelial ovarian cancer after cytoreductive surgery.

BACKGROUND: Mucinous epithelial ovarian cancer (mEOC) is a relatively uncommon subtype of ovarian cancer with special prognostic features, but there is insufficient research in this area. This study aimed to develop a nomogram for the cancer-specific survival (CSS) of mEOC based on Surveillance, Epidemiology, and End Results (SEER) database and externally validate it in National Union of Real World Gynecological Oncology Research and Patient Management (NUWA) platform from China. METHODS: Patients screened from SEER database were allocated into training and internal validation cohort in a ratio of 7: 3, with those from NUWA platform as an external validation cohort. Significant factors selected by Cox proportional hazard regression were applied to establish a nomogram for 3-year and 5-year CSS. The performance of nomogram was assessed by concordance index, calibration curves and Kaplan-Meier (K-M) curves. RESULTS: The training cohort (n = 572) and internal validation cohort (n = 246) were filtered out from SEER database. The external validation cohort contained 186 patients. Baseline age, tumor stage, histopathological grade, lymph node metastasis and residual disease after primary surgery were significant risk factors (p < 0.05) and were included to develop the nomogram. The C-index of nomogram in training, internal validation and external validation cohort were 0.869 (95% confidence interval [CI], 0.838-0.900), 0.839 (95% CI, 0.787-0.891) and 0.800 (95% CI, 0.738-0.862), respectively. The calibration curves of 3-year and 5-year CSS in each cohort showed favorable agreement between prediction and observation. K-M curves of different risk groups displayed great discrimination. CONCLUSION: The discrimination and goodness of fit of the nomogram indicated its satisfactory predictive value for the CSS of mEOC in SEER database and external validation in China, which implies its potential application in different populations.

Platelet RNA enables accurate detection of ovarian cancer: an intercontinental, biomarker identification study.

Platelets are reprogrammed by cancer via a process called education, which favors cancer development. The transcriptional profile of tumor-educated platelets (TEPs) is skewed and therefore practicable for cancer detection. This intercontinental, hospital-based, diagnostic study included 761 treatment-naïve inpatients with histologically confirmed adnexal masses and 167 healthy controls from nine medical centers (China, n = 3; Netherlands, n = 5; Poland, n = 1) between September 2016 and May 2019. The main outcomes were the performance of TEPs and their combination with CA125 in two Chinese (VC1 and VC2) and the European (VC3) validation cohorts collectively and independently. Exploratory outcome was the value of TEPs in public pan-cancer platelet transcriptome datasets. The AUCs for TEPs in the combined validation cohort, VC1, VC2, and VC3 were 0.918 (95% CI 0.889-0.948), 0.923 (0.855-0.990), 0.918 (0.872-0.963), and 0.887 (0.813-0.960), respectively. Combination of TEPs and CA125 demonstrated an AUC of 0.922 (0.889-0.955) in the combined validation cohort; 0.955 (0.912-0.997) in VC1; 0.939 (0.901-0.977) in VC2; 0.917 (0.824-1.000) in VC3. For subgroup analysis, TEPs exhibited an AUC of 0.858, 0.859, and 0.920 to detect early-stage, borderline, non-epithelial diseases and 0.899 to discriminate ovarian cancer from endometriosis. TEPs had robustness, compatibility, and universality for preoperative diagnosis of ovarian cancer since it withstood validations in populations of different ethnicities, heterogeneous histological subtypes, and early-stage ovarian cancer. However, these observations warrant prospective validations in a larger population before clinical utilities.

Does the primary treatment sequence affect post-relapse survival in recurrent epithelial ovarian cancer? A real-world multicentre retrospective study.

OBJECTIVE: To explore the impact of the primary treatment sequence (primary debulking surgery, PDS, versus neoadjuvant chemotherapy and interval debulking surgery, NACT-IDS) on post-relapse survival (PRS) and recurrence characteristics of recurrent epithelial ovarian cancer (REOC). DESIGN: Real-world retrospective study. SETTING: Tertiary hospitals in China. POPULATION: A total of 853 patients with REOC at International Federation of Gynaecology and Obstetrics stages IIIC-IV from September 2007 to June 2020. Overall, 377 and 476 patients received NACT-IDS and PDS, respectively. METHODS: Propensity score-based inverse probability of treatment weighting (IPTW) was performed to balance the between-group differences. MAIN OUTCOME MEASURES: Clinicopathological factors related to PRS. RESULTS: The overall median PRS was 29.3 months (95% CI 27.0-31.5 months). Multivariate analysis before and after IPTW adjustment showed that NACT-IDS and residual R1/R2 disease were independent risk factors for PRS (p < 0.05). Patients with diffuse carcinomatosis and platinum-free interval (PFI) ≤ 12 months had a significantly worse PRS (p < 0.001). Logistic regression analysis revealed that NACT-IDS was an independent risk factor for diffuse carcinomatosis (OR 1.36, 95% CI 1.01-1.82, p = 0.040) and PFI ≤ 12 months (OR 1.59, 95% CI 1.08-2.35, p = 0.019). In IPTW analysis, NACT-IDS was still significantly associated with diffuse carcinomatosis (OR 1.29, 95% CI 1.05-1.58, p = 0.015) and PFI ≤ 12 months (OR 1.90, 95% CI 1.52-2.38, p < 0.001). CONCLUSIONS: The primary treatment sequence may affect the PRS of patients with REOC by altering the recurrence pattern and PFI duration.

The first Chinese National Union of Real-world Gynaecological Oncology Research and Patient Management Platform: A retrospective study.

OBJECTIVE: To produce high-quality, real-world evidence for oncologists by collating scattered gynaecologic oncology (GO) medical records in China. DESIGN: Retrospective study. SETTING: The National Union of Real-world Gynaecological Oncology Research and Patient Management Platform (NUWA platform). SAMPLE: Patient-centred data pool. METHODS: The NUWA platform integrated inpatient/outpatient clinical, gene and follow-up data. Data of 11 456 patients with ovarian cancer (OC) were collected and processed using 91 345 electronic medical records. Structured and unstructured data were de-identified and re-collated into a patient-centred data pool using a predefined GO data model by technology-aided abstraction. MAIN OUTCOME MEASURES: Recent treatment pattern shifts towards precision medicine for OC in China. RESULTS: Thirteen first-tier hospitals across China participated in the NUWA platform up to 7 December 2021. In total, 3504 (30.59%) patients were followed up by a stand-alone patient management centre. The percentage of patients undergoing breast cancer gene (BRCA) mutation tests increased by approximately six-fold between 2017 and 2018. A similar trend was observed in the administration rate of poly(ADP-ribose) polymerase inhibitors as first-line treatment and second-line treatment after September 2018, when olaparib was approved for clinical use in China. CONCLUSION: The NUWA platform has great potential to facilitate clinical studies and support drug development, regulatory reviews and healthcare decision-making.

Oncological big data platforms for promoting digital competencies and professionalism in Chinese medical students: a cross-sectional study.

OBJECTIVES: Advancements in big data technology are reshaping the healthcare system in China. This study aims to explore the role of medical big data in promoting digital competencies and professionalism among Chinese medical students. DESIGN, SETTING AND PARTICIPANTS: This study was conducted among 274 medical students who attended a workshop on medical big data conducted on 8 July 2021 in Tongji Hospital. The workshop was based on the first nationwide multifunction gynecologic oncology medical big data platform in China, at the National Union of Real-World Gynecologic Oncology Research & Patient Management Platform (NUWA platform). OUTCOME MEASURES: Data on knowledge, attitudes towards big data technology and professionalism were collected before and after the workshop. We have measured the four skill categories: doctor‒patient relationship skills, reflective skills, time management and interprofessional relationship skills using the Professionalism Mini-Evaluation Exercise (P-MEX) as a reflection for professionalism. RESULTS: A total of 274 students participated in this workshop and completed all the surveys. Before the workshop, only 27% of them knew the detailed content of medical big data platforms, and 64% knew the potential application of medical big data. The majority of the students believed that big data technology is practical in their clinical practice (77%), medical education (85%) and scientific research (82%). Over 80% of the participants showed positive attitudes toward big data platforms. They also exhibited sufficient professionalism before the workshop. Meanwhile, the workshop significantly promoted students' knowledge of medical big data (p<0.05), and led to more positive attitudes towards big data platforms and higher levels of professionalism. CONCLUSIONS: Chinese medical students have primitive acquaintance and positive attitudes toward big data technology. The NUWA platform-based workshop may potentially promote their understanding of big data and enhance professionalism, according to the self-measured P-MEX scale.

Sintilimab combined with bevacizumab in relapsed/persistent ovarian clear cell carcinoma (INOVA): an investigator-initiated, multicentre clinical trial-a study protocol of clinical trial.

BACKGROUND: Ovarian clear cell carcinoma (OCCC) has an abysmal prognosis with a median overall survival (OS) of 25.3 months because of a low response to chemotherapy. The 5-year disease-specific survival rate after recurrence is 13.2%, with more than two-thirds of the patients dying within a year. Therefore, it is urgent to explore new therapeutic options for OCCC. Based on the characteristic immune-suppressive tumour microenvironment derived from the gene expression profile of OCCC, the combination of immunoantiangiogenesis therapy might have certain efficacy in recurrent/persistent OCCC. This trial aims to evaluate the efficacy and safety of sintilimab and bevacizumab in patients who have failed platinum-containing chemotherapy with recurrent or persistent OCCC. METHOD AND ANALYSIS: In this multicentre, single-arm, open-label, investigator-initiated clinical trial, 38 patients will be assigned to receive sintilimab 200 mg plus bevacizumab 15 mg/kg every 3 weeks. The eligibility criteria include histologically diagnosed patients with recurrent or persistent OCCC who have been previously treated with at least one-line platinum-containing chemotherapy; patients with Eastern Cooperative Oncology Group (ECOG) performance status 0-2 with an expected survival greater than 12 weeks. The exclusion criteria include patients previously treated with immune checkpoint inhibitor and patients with contraindications of bevacizumab and sintilimab. The primary endpoint is the objective response rate. The secondary endpoints are progression-free survival, time to response, duration of response, disease control rate, OS, safety and quality of life. Statistical significance was defined as p<0.05. ETHICS AND DISSEMINATION: This trial was approved by the Research Ethics Commission of Tongji Medical College of Huazhong University of Science and Technology (2020-S337). The protocol of this study is registered at www. CLINICALTRIALS: gov. The trial results will be published in peer-reviewed journals and at conferences. TRIAL REGISTRATION NUMBER: NCT04735861; Clinicaltrials. gov.

Deep learning-enabled pelvic ultrasound images for accurate diagnosis of ovarian cancer in China: a retrospective, multicentre, diagnostic study.

BACKGROUND: Ultrasound is a critical non-invasive test for preoperative diagnosis of ovarian cancer. Deep learning is making advances in image-recognition tasks; therefore, we aimed to develop a deep convolutional neural network (DCNN) model that automates evaluation of ultrasound images and to facilitate a more accurate diagnosis of ovarian cancer than existing methods. METHODS: In this retrospective, multicentre, diagnostic study, we collected pelvic ultrasound images from ten hospitals across China between September 2003, and May 2019. We included consecutive adult patients (aged ≥18 years) with adnexal lesions in ultrasonography and healthy controls and excluded duplicated cases and patients without adnexa or pathological diagnosis. For DCNN model development, patients were assigned to the training dataset (34 488 images of 3755 patients with ovarian cancer, 541 442 images of 101 777 controls). For model validation, patients were assigned to the internal validation dataset (3031 images of 266 patients with ovarian cancer, 5385 images of 602 with benign adnexal lesions), external validation datasets 1 (486 images of 67 with ovarian cancer, 933 images of 268 with benign adnexal lesions), and 2 (1253 images of 166 with ovarian cancer, 5257 images of 723 benign adnexal lesions). Using these datasets, we assessed the diagnostic value of DCNN, compared DCNN with 35 radiologists, and explored whether DCNN could augment the diagnostic accuracy of six radiologists. Pathological diagnosis was the reference standard. FINDINGS: For DCNN to detect ovarian cancer, AUC was 0·911 (95% CI 0·886-0·936) in the internal dataset, 0·870 (95% CI 0·822-0·918) in external validation dataset 1, and 0·831 (95% CI 0·793-0·869) in external validation dataset 2. The DCNN model was more accurate than radiologists at detecting ovarian cancer in the internal dataset (88·8% vs 85·7%) and external validation dataset 1 (86·9% vs 81·1%). Accuracy and sensitivity of diagnosis increased more after DCNN-assisted diagnosis than assessment by radiologists alone (87·6% [85·0-90·2] vs 78·3% [72·1-84·5], p<0·0001; 82·7% [78·5-86·9] vs 70·4% [59·1-81·7], p<0·0001). The average accuracy of DCNN-assisted evaluations for six radiologists reached 0·876 and were significantly augmented when they were DCNN-assisted (p<0·05). INTERPRETATION: The performance of DCNN-enabled ultrasound exceeded the average diagnostic level of radiologists matched the level of expert ultrasound image readers, and augmented radiologists' accuracy. However, these observations warrant further investigations in prospective studies or randomised clinical trials. FUNDING: National Key Basic Research Program of China, National Sci-Tech Support Projects, and National Natural Science Foundation of China.

PRCTC: a machine learning model for prediction of response to corticosteroid therapy in COVID-19 patients.

Corticosteroid has been proved to be one of the few effective treatments for COVID-19 patients. However, not all the patients were suitable for corticosteroid therapy. In this study, we aimed to propose a machine learning model to forecast the response to corticosteroid therapy in COVID-19 patients. We retrospectively collected the clinical data about 666 COVID-19 patients receiving corticosteroid therapy between January 27, 2020, and March 30, 2020, from two hospitals in China. The response to corticosteroid therapy was evaluated by hospitalization time, oxygen supply duration, and the outcomes of patients. Least Absolute Shrinkage and Selection Operator (LASSO) was applied for feature selection. Five prediction models were applied in the training cohort and assessed in an internal and an external validation dataset, respectively. Finally, two (C reactive protein, lymphocyte percent) of 36 candidate immune/inflammatory features were finally used for model development. All five models displayed promising predictive performance. Notably, the ensemble model, PRCTC (prediction of response to corticosteroid therapy in COVID-19 patients), derived from three prediction models including Gradient Boosted Decision Tree (GBDT), Neural Network (NN), and logistic regression (LR), achieved the best performance with an area under the curve (AUC) of 0.810 (95% confidence interval [CI] 0.760-0.861) in internal validation cohort and 0.845 (95% CI 0.779-0.911) in external validation cohort to predict patients' response to corticosteroid therapy. In conclusion, PRCTC proposed with universality and scalability is hopeful to provide tangible and prompt clinical decision support in management of COVID-19 patients and potentially extends to other medication predictions.

Prevalence and risk factors of thrombotic events on patients with COVID-19: a systematic review and meta-analysis.

BACKGROUND: Coagulation abnormalities in COVID-19 patients accompanied with poor prognosis. This study aimed to determine the prevalence and risk factors of thrombotic events on COVID-19 patients. METHODS: We systematically reviewed all the studies about thrombotic events on COVID-19 patients in PubMed, Embase, Web of Science, MedRxiv, bioRxiv, from Dec 1, 2019 to July 5, 2020. The weighted mean difference (MD) or odds ratio (OR) or relative risk (RR) with 95 % confidence intervals (CI) for clinical data in COVID-19 patients with or without thrombotic events was calculated. RESULTS: 12 articles contained 1083 patients were included for meta-analysis. The prevalence of thrombosis was 22 % (95 % CI 0.08-0.40) in COVID-19 patients and increased to 43 % (95 % CI 0.29-0.65) after admission to the intensive care unit (ICU). Compared with non-thrombotic patients, thrombotic patients had higher levels of D-dimer (MD = 2.79 µg/ml, 95 % CI 2.27-3.31 µg/ml), lactate dehydrogenase (LDH) (MD = 112.71 U/L, 95 % CI 62.40-163.02 U/L), and white blood cells (WBC) (MD = 1.14 *109/L, 95 % CI 0.47-1.81*109/L) while decreased lymphocytes (MD= -0.20*109/L, 95 % CI -0.38 - -0.02*109/L). Age, platelet counts, and male sex tended to be risks while diabetes tended to be a protection for thrombosis for COVID-19 patients, although no statistical difference was achieved. Finally, patients with thrombosis were at a higher risk of death (OR = 2.39, 95 % CI 1.36-4.20). CONCLUSIONS: Prevalence of thrombosis in COVID-19 patients was high, especially in ICU, though pharmacologic thromboembolism prophylaxis was applied. Therefore, higher levels of D-dimer, LDH, WBC, and decreased lymphocytes needed to be paid close attention to in patients with COVID-19.

Development and validation of an online model to predict critical COVID-19 with immune-inflammatory parameters.

BACKGROUND: Immune and inflammatory dysfunction was reported to underpin critical COVID-19(coronavirus disease 2019). We aim to develop a machine learning model that enables accurate prediction of critical COVID-19 using immune-inflammatory features at admission. METHODS: We retrospectively collected 2076 consecutive COVID-19 patients with definite outcomes (discharge or death) between January 27, 2020 and March 30, 2020 from two hospitals in China. Critical illness was defined as admission to intensive care unit, receiving invasive ventilation, or death. Least Absolute Shrinkage and Selection Operator (LASSO) was applied for feature selection. Five machine learning algorithms, including Logistic Regression (LR), Support Vector Machine (SVM), Gradient Boosted Decision Tree (GBDT), K-Nearest Neighbor (KNN), and Neural Network (NN) were built in a training dataset, and assessed in an internal validation dataset and an external validation dataset. RESULTS: Six features (procalcitonin, [T + B + NK cell] count, interleukin 6, C reactive protein, interleukin 2 receptor, T-helper lymphocyte/T-suppressor lymphocyte) were finally used for model development. Five models displayed varying but all promising predictive performance. Notably, the ensemble model, SPMCIIP (severity prediction model for COVID-19 by immune-inflammatory parameters), derived from three contributive algorithms (SVM, GBDT, and NN) achieved the best performance with an area under the curve (AUC) of 0.991 (95% confidence interval [CI] 0.979-1.000) in internal validation cohort and 0.999 (95% CI 0.998-1.000) in external validation cohort to identify patients with critical COVID-19. SPMCIIP could accurately and expeditiously predict the occurrence of critical COVID-19 approximately 20 days in advance. CONCLUSIONS: The developed online prediction model SPMCIIP is hopeful to facilitate intensive monitoring and early intervention of high risk of critical illness in COVID-19 patients. TRIAL REGISTRATION: This study was retrospectively registered in the Chinese Clinical Trial Registry ( ChiCTR2000032161 ). vv.

Clinical characteristics and outcomes of pregnant women with COVID-19 and the risk of vertical transmission: a systematic review.

PURPOSE: This systematic review summarizes the clinical features and maternal-infant outcomes of 230 pregnant women (154 patients gave birth) infected with COVID-19 and their 156 infants, including the possibility and evidence of vertical transmission. METHODS: An electronic search of PubMed, Embase, Medline, MedRxiv, CNKI, and the Chinese Medical Journal Full Text Database following PRISMA guidelines was performed through April 18, 2020. Search terms included COVID-19, SARS-CoV-2, pregnant women, infants, and vertical transmission. RESULTS: A total of 230 women with COVID-19 (154 deliveries, 66 ongoing pregnancies, and 10 abortions) and 156 newborns from 20 eligible studies were included in this systematic review. A total of 34.62% of the pregnant patients had obstetric complications, and 59.05% of patients displayed fever. Lymphopenia was observed in 40.71% of patients. A total of 5.19% of women received mechanical ventilation. Seven women were critically ill. One mother and two newborns died. A total of 24.74% of newborns were premature. Five newborns' throat swab tests of SARS-CoV-2 were positive, all of which were delivered by cesarean section. For eight newborns with negative throat swab tests, three had both elevated IgM and IgG against SARS-CoV-2. Nucleic acid tests of vaginal secretions, breast milk, amniotic fluid, placental blood, and placental tissues were negative. CONCLUSION: Most pregnant patients were mildly ill. The mortality of pregnant women with COVID-19 was lower than that of overall COVID-19 patients. Cesarean section was more common than vaginal delivery for pregnant women with COVID-19. Premature delivery was the main adverse event for newborns. The vertical transmission rate calculated by SARS-CoV-2 nucleic acid tests was 3.91%. Serum antibodies against SARS-CoV-2 should be tested more frequently, and multiple samples should be included in pathogenic testing.

Machine learning based early warning system enables accurate mortality risk prediction for COVID-19.

Soaring cases of coronavirus disease (COVID-19) are pummeling the global health system. Overwhelmed health facilities have endeavored to mitigate the pandemic, but mortality of COVID-19 continues to increase. Here, we present a mortality risk prediction model for COVID-19 (MRPMC) that uses patients' clinical data on admission to stratify patients by mortality risk, which enables prediction of physiological deterioration and death up to 20 days in advance. This ensemble model is built using four machine learning methods including Logistic Regression, Support Vector Machine, Gradient Boosted Decision Tree, and Neural Network. We validate MRPMC in an internal validation cohort and two external validation cohorts, where it achieves an AUC of 0.9621 (95% CI: 0.9464-0.9778), 0.9760 (0.9613-0.9906), and 0.9246 (0.8763-0.9729), respectively. This model enables expeditious and accurate mortality risk stratification of patients with COVID-19, and potentially facilitates more responsive health systems that are conducive to high risk COVID-19 patients.

Risk factors for developing into critical COVID-19 patients in Wuhan, China: A multicenter, retrospective, cohort study.

BACKGROUND: The ferocious global assault of COVID-19 continues. Critically ill patients witnessed significantly higher mortality than severe and moderate ones. Herein, we aim to comprehensively delineate clinical features of COVID-19 and explore risk factors of developing critical disease. METHODS: This is a Mini-national multicenter, retrospective, cohort study involving 2,387 consecutive COVID-19 inpatients that underwent discharge or death between January 27 and March 21, 2020. After quality control, 2,044 COVID-19 inpatients were enrolled. Electronic medical records were collected to identify the risk factors of developing critical COVID-19. FINDINGS: The severity of COVID-19 climbed up straightly with age. Critical group was characterized by higher proportion of dyspnea, systemic organ damage, and long-lasting inflammatory storm. All-cause mortality of critical group was 85•45%, by contrast with 0•58% for severe group and 0•18% for moderate group. Logistic regression revealed that sex was an effect modifier for hypertension and coronary heart disease (CHD), where hypertension and CHD were risk factors solely in males. Multivariable regression showed increasing odds of critical illness associated with hypertension, CHD, tumor, and age ≥ 60 years for male, and chronic obstructive pulmonary disease (COPD), chronic kidney disease (CKD), tumor, and age ≥ 60 years for female. INTERPRETATION: We provide comprehensive front-line information about different severity of COVID-19 and insights into different risk factors associated with critical COVID-19 between sexes. These results highlight the significance of dividing risk factors between sexes in clinical and epidemiologic works of COVID-19, and perhaps other coronavirus appearing in future. FUNDING: 10.13039/100000001 National Science Foundation of China.