Search

Considering difficulties in on-site ADAMTS13 testing and the performance instability of PLASMIC score according to ethnicity, we developed a prediction tool, MED-TMA (machine learning (ML) method for differential diagnosis (DDx) of thrombotic microangiopathy (TMA)) to support clinical decision. Data from 319 patients visiting 31 hospitals in Korea clinically diagnosed with primary TMA was randomly separated by 2:1 into two groups – the development dataset (D-set, n = 212), the validation dataset (V-set, n = 107). Feature elimination was conducted to select optimal clinical predictors. We developed the model with the selected features using ML methods, verifying using V-set. After the feature elimination using 19 clinical variables, five variables were selected with high importance value. Among nine ML methods, four ML methods were chosen considering the Area Under the Curves (AUC) and the correlation between the methods using D-set. We developed MED-TMA based on an optimized ensemble model with the selected four ML methods resulting in AUC values of 0.945 and 0.924 in D-set and V-set, respectively. In addition to the binary outcome, MED-TMA was capable of providing a probability for DDx of TMA. The ensemble approach driven MED-TMA showed comparable accurate and intuitive decision support for DDx of TMA to that of the existing models based on a single ML method. We provide a web-based nomogram for the appropriate use of effective but costly therapeutics to treat TMA patients (http://hematology.snu.ac.kr/medtma/). • For differential diagnosis of primary TMA, we propose a decision support tool named MED-TMA. • MED-TMA improves the performance through ensemble with various ML methods. • MED-TMA provides accurate results with using simple laboratory data of patients. • MED-TMA provides more intuitive suggestion compared with the existing models. [ABSTRACT FROM AUTHOR]

To compare survival outcomes of primary laparoscopic radical hysterectomy (LRH) and open radical hysterectomy (ORH) in patients with FIGO stage IB cervical cancer. We retrospectively identified stage IB1–IB2 cervical cancer patients who received either LRH (n = 343) or ORH (n = 222) at two tertiary institutional hospitals between 2000 and 2018. To adjust for confounders, we conducted Mahalanobis distance-based sample matching for stage, histology, cervical mass size, parametrial invasion, and lymph node metastasis. Then, survival outcomes were compared between the matched groups. Through the independent matching processes, we narrowed the study population to stage IB1 patients and stage IB1 patients with tumor size ≤2 cm on pre-operative MRI. After matching, LRH group showed poorer progression-free survival (PFS) than ORH group (3-year: 85.4% vs. 91.8%; P = 0.036), whereas no significant difference in overall survival (OS) was found. Regarding recurrence patterns, no significant differences in the incidences of pelvic, retroperitoneal lymph node and abdominal recurrences, or distant metastasis were observed between the two groups. Among the matched patients with stage IB1 who had cervical mass size ≤2 cm, the LRH and ORH groups showed similar PFS (3-year: 90.0% vs. 93.1%; P = 0.8) and OS (5-year: 98.6% vs. 96.4%; P = 0.6). Despite the retrospective design, our matched cohort study suggests that ORH might be preferable for the surgical treatment of FIGO stage IB cervical cancer. However, in stage IB1 patients with tumor size ≤2 cm, LRH might be applicable, as equivalent outcomes were found regardless of the surgical approach. Further prospective studies are warranted. • We conducted a matching study to investigate survival of laparoscopic radical hysterectomy (LRH) in stage IB cervical cancer. • After matching, LRH group showed a higher recurrence rate compared to open group, whereas overall survival was not different. • In stage IB1 patients with tumor size ≤2 cm, equivalent survival outcomes were observed regardless of the surgical approach. [ABSTRACT FROM AUTHOR]

Background: Technologically integrated healthcare environments can be realized if physicians are encouraged to use smart systems for the creation and sharing of knowledge used in clinical decision support systems (CDSS). Physicians are challenged to personalize care with rapidly changing scientific evidence, drug approvals, and guidelines for heart failure (HF). Artificial intelligence (AI) CDSS have the potential to help lessen this challenge. Hypothesis: We report here the results of examining the level of correspondence between diagnostic recommendations made by the AI CDSS and a cardiology board for HF.. Method: AI CDSS system for cardiology (e-Health for cardiology (eHFC)) was developed by using Intelligence Knowledge Authoring Tool (I-KAT) and archiving Arden Syntax MLM (Medical Logic Module) as shareable knowledge rules for intelligent decision-making. A total of 598 patients with suspected HF were available between 2016 and 2017 at Seoul National University Bundang Hospital, South Korea. Both eHFC and HF specialist gave a diagnosis for each patient independently. Concordance rate between eHFC and HF specialist was evaluated. Results: Overall, the concordance rate was 98.3%. The concordance rate for patients with HFrEF, HFmrEF, HFpEF and no-HF was 100%, 100%, 99.56%, and 91.67%, respectively (Table 1): 9 patients without HF were wrongly classified as HFpEF. The concordance was independent of age of the patients. We divided patients according to echocardiographic parameters. Set A (all echocardiography parameters were available) and Set B (only LVEF, LAVI, and LVMI were available). The concordance rate was lower in Set B than in Set A (Figure). Conclusions: eHFC shows high diagnostic accuracy to diagnose HF, independent of HF types. eHFC may be useful for diagnosis of HF, especially at centers without HF specialist. Key words: artificial intelligence, clinical decision-support system, heart failure Table 1: Confusion Matrix of eHFC for HFrEF, HFmrEF, HFpEF, and no-HF Figure 1: Accuracy Comparison of Set A, and Set B [ABSTRACT FROM AUTHOR]