Taiichi Wakiya

Lecture

Jichi Medical University
Country flagJapan

Research Interests

ApplyKite TwinLearns as you apply

Make your next move.

Prepare your application

Review drafts before you send them.

Explore related searches

Articles (20)

Year:

2026

Year:

2026

Background Massive intraoperative bleeding (IBL) in liver transplantation (LT) poses serious risks and strains healthcare resources necessitating better predictive models for risk stratification. As traditional models often fail to capture the complex, non-linear patterns underlying bleeding risk, this study aimed to develop data-driven machine learning models for predicting massive IBL during living donor LT (LDLT) using preoperative factors. Methods Two hundred ninety consecutive LDLT cases from a prospective database were analyzed. Logistic regression models were built using 73 preoperative demographic and laboratory variables to predict massive IBL (≥ 80 mL/kg). The dataset was randomly split (70% training, 30% testing). The model was trained and validated through three-fold cross-validation, with backward stepwise feature selection iterated 100 times across unique random splits. The final model, based on a high stability index, was evaluated using the area under the curve (AUC). Results Massive IBL was observed in 141 patients (48.6%). In standard logistic regression, significant differences were found in 42 of 73 factors between groups stratified by massive IBL, however, substantial multicollinearity limited interpretability. In the feature selection across 100 iterations, the data-driven model achieved an average AUC of 0.840 in the validation and 0.738 in the test datasets. The final model, based on 11 selected features with a high stability index, achieved an AUC of 0.844. An easy-to-use online risk calculator for massive IBL was developed and is available at: https://tai1wakiya.shinyapps.io/ldlt_bleeding_ml/ . Conclusions Our findings highlight the potential of machine learning in capturing complex risk factor interactions for predicting massive IBL in LDLT.

Year:

2026

Year:

2025

While endogenous cortisol secretion rises in the early morning, the number of lymphocytes in the blood is higher at night, thus exhibiting an antiphase pattern to cortisol secretion. Therefore, compared with the daytime, the infiltration of lymphocytes into immune‐reactive tissues is enhanced at night. This study aimed to determine whether the administration of methylprednisolone (mPSL) in the evening is more effective against T cell‐mediated rejection (TCMR) after liver transplantation compared with morning administration. This study used a randomized, open‐label, parallel‐group comparison design. Pediatric patients scheduled to undergo living‐donor liver transplantation were randomly divided into morning (8:00 a.m.) and evening (8:00 p.m.) mPSL administration groups. The primary endpoint was the occurrence of TCMR within 14 days of surgery. Sixty‐two patients were enrolled between 2014 and 2023, and six patients were excluded from the analysis as their dose of mPSL deviated from the protocol within 14 days after surgery. Of the 56 subjects analyzed, TCMR was detected in 10 of the morning group ( n = 29) and three of the evening group ( n = 27) within 14 days after surgery. Stratified analysis of patients who did not receive preoperative rituximab treatment showed that none of the evening group and 36.4% of the morning group developed TCMR within 14 days after surgery ( P < 0.01, 95% confidence interval; 2.00–infinity). Safety evaluation results were comparable between the two groups. This study shows that the evening administration of mPSL is an effective approach for suppressing TCMR. This study is hypothesis generating, and replication in further studies is needed.

Year:

2025

Background Living donor hepatectomy carries a risk of bleeding, and allogeneic transfusion may cause adverse effects. Autologous blood preparation is therefore considered a safer strategy. While preoperative autologous donation (PAD) is widely used, acute normovolemic hemodilution (ANH) may overcome its limitations and offer a practical alternative. However, evidence supporting ANH in donor surgery remains limited. This study compared the clinical utility of ANH and PAD in living donor hepatectomy. Methods We retrospectively analyzed 60 consecutive cases of living donor right hepatectomy performed between 2017 and 2025. Among them, 58 donors who received either PAD or ANH were compared using 1:2 propensity score matching. Perioperative laboratory values, surgical outcomes, any allogeneic transfusion, and postoperative complications were evaluated. Results No unexpected intraoperative adverse events or allogeneic transfusions occurred. All ANH donors received autologous reinfusion, compared with only 6.5% of PAD donors. After matching, 27 donors (18 PAD, 9 ANH) were analyzed. Whole blood viscosity was higher in the ANH group. Weight‐adjusted intraoperative bleeding was lower (2.6 vs. 4.6 mL/kg, p = .024; q = 0.106) and operative time was shorter (321 vs. 390 min, p = .007; q = 0.077) in the ANH group. Postoperative complication rates were comparable. Area under the curve analysis indicated better preservation of total protein ( p = .038) and prothrombin time‐international normalized ratio ( p = .010) across the perioperative period in the ANH group. Conclusions ANH maintained transfusion avoidance and donor safety comparable to PAD while improving operative efficiency, supporting ANH as a safe, effective alternative in living donor right hepatectomy.

Year:

2025

Year:

2025

Background Massive intraoperative bleeding (IBL) in liver transplantation (LT) poses serious risks and strains healthcare resources necessitating better predictive models for risk stratification. As traditional models often fail to capture the complex, non-linear patterns underlying bleeding risk, this study aimed to develop data-driven machine learning models for predicting massive IBL during LT using preoperative factors. Methods Two hundred ninety consecutive LT cases from a prospective database were analyzed. Logistic regression models were built using 73 preoperative demographic and laboratory variables to predict massive IBL (≥ 80 mL/kg). The dataset was randomly split (70% training, 30% testing). The model was trained and validated through three-fold cross-validation, with backward stepwise feature selection iterated 100 times across unique random splits. The final model, based on a high stability index, was evaluated using the area under the curve (AUC). Results Massive IBL was observed in 141 patients (48.6%). In standard logistic regression, significant differences were found in 42 of 73 factors between groups stratified by massive IBL, however, substantial multicollinearity limited interpretability. In the feature selection across 100 iterations, the data-driven model achieved an average AUC of 0.840 in the validation and 0.738 in the test datasets. The final model, based on 11 selected features with a high stability index, achieved an AUC of 0.844. An easy-to-use online risk calculator for massive IBL was developed and is available at: https://tai1wakiya.shinyapps.io/ldlt_bleeding_ml/ . Conclusions Our findings highlight the potential of machine learning in capturing complex risk factor interactions for predicting massive IBL in LT.

Year:

2025

Year:

2024

Year:

2024

Year:

2024

Year:

2024

Year:

2024

Year:

2024

Year:

2024

Year:

2024

Year:

2023

Year:

2023

Year:

2023

Year:

2023

Collaborators (6)

Seiichi Kato

Professor

Saga University

JAPAN
View details

Kentaro Ushijima

Sanyo-Onoda City University

JAPAN
View details

Shinya Okumura

Assistant Professor

Kyoto University Hospital

JAPAN
View details

Hideo Ogiso

-

JAPAN
View details

Masaki Honda

Kumamoto University Hospital

JAPAN
View details

Norihisa Kimura

Lecturer

Hirosaki University Graduate School of Medicine

JAPAN
View details