Annisa Ristya Rahmanti
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Annisa Ristya Rahmanti has a total of 67 co-authors. Here are the top co-authors by number of shared publications:
- Yu-Chuan (Jack) Li
Graduate Institute of Biomedical Informatics, College of Medical Science & Technology, Taipei Medical University, Taipei, Taiwan
9 shared publications - Hsuan-Chia Yang
Graduate Institute of Biomedical Informatics, College of Medical Science and Technology, Taipei Medical University, Taipei, Taiwan
7 shared publications - Chih-Wei Huang
International Center for Health Information Technology, College of Medical Science and Technology, Taipei Medical University, Taipei, Taiwan
5 shared publications Lutfan Lazuardi
Universitas Gadjah Mada
4 shared publications- Huan Xuan Nguyen
Faculty of Science and Technology, Middlesex University, London NW4 4BT, UK
3 shared publications - Xiaohong Gao
Middlesex University
3 shared publications
Latest Publications
- 17 September 2026
- Published by Elsevier in International Journal of Medical Informatics
- Vol. 223
Machine learning for risk prediction of infectious diseases: a scoping review of application, predictive performance, and reporting gaps
- 17 September 2026
- Published by Elsevier in International Journal of Medical Informatics
- Vol. 223
Machine learning for risk prediction of infectious diseases: a scoping review of application, predictive performance, and reporting gaps
- Open Access
- 15 June 2026
- Published by Springer Nature in SN Computer Science
- Vol. 7  (5)
Neural Radiance Field Based 3D View Reconstruction for Gastrointestinal Tract Surgery Planning
Gastrointestinal (GI) tract is a 9-meter-long food passage that transports food and nutrition to the human body. Due to the narrow space (~ 2 cm in diameter) of the GI tract, a lesion removal surgery through endoscopic procedures faces limited view of the lesion, leading to the challenge of maximum removal of diseased lesion and minimum sacrifice of healthy tissues. This paper evaluates the application of the state-of-the-art AI technique of neural radiance field to reconstruct 3D models of a lesion based on 2D endoscopic videos. In this way, a full 3D view of concerned GI lesion can be presented to surgeons to allow comprehensive surgical planning. The advanced AI techniques, neural radiance field (NeRF), structure-from-motion and multi-view stereo are applied. This system is implemented based on COLMAP and Nerfstudio libraries. The system is evaluated using two sets of video clips containing 2600 images. Initial results illustrate that this end-to-end deep learning architecture, i.e. from 2D video input to 3D model output, presents considerable potential for reconstruction of GI lesions. The similarity measures of SSIM, PSNR and LPIPS between original (ground truth) and rendered images are of 19.46 ± 2.56, 0.70 ± 0.054, and 0.49 ± 0.05 respectively. This work contributes to the solution towards one of the long-standing challenges in computer vision field, which is to construct 3D views from 2D videos in the field of GI surgery planning and has achieved a promising performance. Future work includes enlarging datasets and removal of ghostly artefact from rendered images.
- 15 June 2026
- Published by Springer Nature in SN Computer Science
- Vol. 7  (5)
Open AccessNeural Radiance Field Based 3D View Reconstruction for Gastrointestinal Tract Surgery Planning
Gastrointestinal (GI) tract is a 9-meter-long food passage that transports food and nutrition to the human body. Due to the narrow space (~ 2 cm in diameter) of the GI tract, a lesion removal surgery through endoscopic procedures faces limited view of the lesion, leading to the challenge of maximum removal of diseased lesion and minimum sacrifice of healthy tissues. This paper evaluates the application of the state-of-the-art AI technique of neural radiance field to reconstruct 3D models of a lesion based on 2D endoscopic videos. In this way, a full 3D view of concerned GI lesion can be presented to surgeons to allow comprehensive surgical planning. The advanced AI techniques, neural radiance field (NeRF), structure-from-motion and multi-view stereo are applied. This system is implemented based on COLMAP and Nerfstudio libraries. The system is evaluated using two sets of video clips containing 2600 images. Initial results illustrate that this end-to-end deep learning architecture, i.e. from 2D video input to 3D model output, presents considerable potential for reconstruction of GI lesions. The similarity measures of SSIM, PSNR and LPIPS between original (ground truth) and rendered images are of 19.46 ± 2.56, 0.70 ± 0.054, and 0.49 ± 0.05 respectively. This work contributes to the solution towards one of the long-standing challenges in computer vision field, which is to construct 3D views from 2D videos in the field of GI surgery planning and has achieved a promising performance. Future work includes enlarging datasets and removal of ghostly artefact from rendered images.
- Open Access
- 21 May 2026
- Published by IOS Press
- Vol. 336, 218-222
Prediction of Early Functional Outcome After Acute Ischemic Stroke Using Real-World Clinical Data in Vietnam and Indonesia: Retrospective Cohort Study
Accurate prediction of early functional outcome after acute ischemic stroke is critical for clinical decision-making. This retrospective cohort study developed and externally validated a machine learning model using routine clinical data from two settings. A total of 11,911 ischemic patients from the 2023 Stroke Care Quality (RES-Q) registry in 52 hospitals across Vietnam and 83 patients from UGM Academic Hospital, Indonesia, were included. The primary outcome was discharge modified Rankin Scale (mRS≤2=favorable; >2 = poor). The XGBoost model achieved strong internal discrimination (AUC=0.905, F1=0.836) and maintained robust external performance (AUC=0.888, F1=0.810). SHAP interpretability identified pre-stroke mRS, admission NIHSS score, first-day glucose check, and age as the strongest predictors of poor functional outcome, while early dysphagia screening, physiotherapy evaluation, and small-vessel etiology were associated with better recovery. These results demonstrate that a data-driven AI approach using routine clinical parameters can achieve reliable cross-regional generalization and may support early prognostication in diverse stroke care settings.
- 21 May 2026
- Published by IOS Press
- Vol. 336, 218-222
Open AccessPrediction of Early Functional Outcome After Acute Ischemic Stroke Using Real-World Clinical Data in Vietnam and Indonesia: Retrospective Cohort Study
Accurate prediction of early functional outcome after acute ischemic stroke is critical for clinical decision-making. This retrospective cohort study developed and externally validated a machine learning model using routine clinical data from two settings. A total of 11,911 ischemic patients from the 2023 Stroke Care Quality (RES-Q) registry in 52 hospitals across Vietnam and 83 patients from UGM Academic Hospital, Indonesia, were included. The primary outcome was discharge modified Rankin Scale (mRS≤2=favorable; >2 = poor). The XGBoost model achieved strong internal discrimination (AUC=0.905, F1=0.836) and maintained robust external performance (AUC=0.888, F1=0.810). SHAP interpretability identified pre-stroke mRS, admission NIHSS score, first-day glucose check, and age as the strongest predictors of poor functional outcome, while early dysphagia screening, physiotherapy evaluation, and small-vessel etiology were associated with better recovery. These results demonstrate that a data-driven AI approach using routine clinical parameters can achieve reliable cross-regional generalization and may support early prognostication in diverse stroke care settings.
- Open Access
- 21 May 2026
- Published by IOS Press
- Vol. 336, 358-362
Machine Learning Applications Within the Earlier Medicine Framework for Stroke: A Scoping Review
This scoping review explores how machine learning (ML) has been applied to stroke research within the Earlier Medicine framework, which promotes proactive and personalized care through primary (preventive care), secondary (acute care), and tertiary (chronic care) prevention levels. Following PRISMA-ScR guidelines, studies were identified from PubMed and Scopus, yielding 105 eligible studies. Most studies focused on tertiary prevention, such as predicting in-hospital mortality, complications, and functional outcomes, while fewer addressed early risk prediction or acute-phase detection. Common predictors included age, NIHSS score, glucose, stroke volume, BMI, hypertension, and diabetes, reflecting reliance on routine clinical data. Traditional ML models (e.g., logistic regression, random forest, SVM) remain dominant, although Ensemble/Hybrid and Deep Learning models have increased steadily since 2021. ML demonstrated promising predictive accuracy and clinical potential, though challenges remain in data heterogeneity, model transparency, and external validation.
- 21 May 2026
- Published by IOS Press
- Vol. 336, 358-362
Open AccessMachine Learning Applications Within the Earlier Medicine Framework for Stroke: A Scoping Review
This scoping review explores how machine learning (ML) has been applied to stroke research within the Earlier Medicine framework, which promotes proactive and personalized care through primary (preventive care), secondary (acute care), and tertiary (chronic care) prevention levels. Following PRISMA-ScR guidelines, studies were identified from PubMed and Scopus, yielding 105 eligible studies. Most studies focused on tertiary prevention, such as predicting in-hospital mortality, complications, and functional outcomes, while fewer addressed early risk prediction or acute-phase detection. Common predictors included age, NIHSS score, glucose, stroke volume, BMI, hypertension, and diabetes, reflecting reliance on routine clinical data. Traditional ML models (e.g., logistic regression, random forest, SVM) remain dominant, although Ensemble/Hybrid and Deep Learning models have increased steadily since 2021. ML demonstrated promising predictive accuracy and clinical potential, though challenges remain in data heterogeneity, model transparency, and external validation.
- Open Access
- 7 August 2025
- Published by IOS Press
- Vol. 329, 1327-1331
Design and Simulation-Based Testing of a Mobile Application for Self-Care Management in Type 2 Diabetes Mellitus and Prediabetes
Mobile health applications are increasingly recognized as valuable tools for supporting self-care management among individuals with type 2 diabetes mellitus (T2DM) and prediabetes. This study explores the development and testing of NusaCare, a mobile app designed to assist patients through lifestyle tracking, education, and personalized interventions. Using a co-design approach, the process began with desk reviews and focus group discussions to gather initial insights. A pre-simulation phase followed, during which prototypes were tested and refined. The final prototype was evaluated in a simulation involving 61 participants, revealing that the challenges feature (70%) and food tracking (42.86%) were the most frequently used features. Post-intervention feedback highlighted the app’s usability and its potential to enhance self-care behaviors. This study emphasizes the importance of engaging users in co-design processes to develop effective mobile health interventions for managing prediabetes and T2DM.
- 7 August 2025
- Published by IOS Press
- Vol. 329, 1327-1331
Open AccessDesign and Simulation-Based Testing of a Mobile Application for Self-Care Management in Type 2 Diabetes Mellitus and Prediabetes
Mobile health applications are increasingly recognized as valuable tools for supporting self-care management among individuals with type 2 diabetes mellitus (T2DM) and prediabetes. This study explores the development and testing of NusaCare, a mobile app designed to assist patients through lifestyle tracking, education, and personalized interventions. Using a co-design approach, the process began with desk reviews and focus group discussions to gather initial insights. A pre-simulation phase followed, during which prototypes were tested and refined. The final prototype was evaluated in a simulation involving 61 participants, revealing that the challenges feature (70%) and food tracking (42.86%) were the most frequently used features. Post-intervention feedback highlighted the app’s usability and its potential to enhance self-care behaviors. This study emphasizes the importance of engaging users in co-design processes to develop effective mobile health interventions for managing prediabetes and T2DM.
- Open Access
- 15 May 2025
- Published by IOS Press
- Vol. 327, 843-847
ChatEndoscopist: A Domain-Specific Chatbot with Images for Gastrointestinal Diseases
This study aims to enhance domain-specific medical knowledge within large language models (LLMs) by developing a chatbot, chatEndoscopist, a specialized model for oesophageal cancer. In particular, the chatbot incorporates related images to further elucidate the retrieved content while providing answers. Fine-tuned BioMistral LLM with 50 related documents, a dataset specifically curated for medical literature, ChatEndoscopist was compared to ChatGPT. For text answers, despite its specialized training, ChatGPT appears to outperform ChatEndoscopist in precision (0.210 vs. 0.148), recall (0.323 vs. 0.049), and F1 score (0.266 vs. 0.099). ChatGPT also demonstrated superior lexical diversity with a Type-Token Ratio (TTR) of 0.772 and Lexical Density of 0.813, compared to ChatEndoscopist’s TTR of 0.717 and Lexical Density of 0.781. This in part, could be due to the limited documents to fine tune. However, the related images are mostly retrieval with regarding to user’s queries. Future work will focus on incorporating more related papers to balance specialized accuracy with broader linguistic flexibility.
- 15 May 2025
- Published by IOS Press
- Vol. 327, 843-847
Open AccessChatEndoscopist: A Domain-Specific Chatbot with Images for Gastrointestinal Diseases
This study aims to enhance domain-specific medical knowledge within large language models (LLMs) by developing a chatbot, chatEndoscopist, a specialized model for oesophageal cancer. In particular, the chatbot incorporates related images to further elucidate the retrieved content while providing answers. Fine-tuned BioMistral LLM with 50 related documents, a dataset specifically curated for medical literature, ChatEndoscopist was compared to ChatGPT. For text answers, despite its specialized training, ChatGPT appears to outperform ChatEndoscopist in precision (0.210 vs. 0.148), recall (0.323 vs. 0.049), and F1 score (0.266 vs. 0.099). ChatGPT also demonstrated superior lexical diversity with a Type-Token Ratio (TTR) of 0.772 and Lexical Density of 0.813, compared to ChatEndoscopist’s TTR of 0.717 and Lexical Density of 0.781. This in part, could be due to the limited documents to fine tune. However, the related images are mostly retrieval with regarding to user’s queries. Future work will focus on incorporating more related papers to balance specialized accuracy with broader linguistic flexibility.
- Open Access
- 8 May 2025
- Published by Taylor & Francis in Medical Education Online
- Vol. 30  (1) , 2497328
Validating nonverbal cues for assessing physician empathy in telemedicine: a Delphi study
Nonverbal communication is essential in physician-patient interaction, especially in telemedicine where verbal cues may be limited. This study aimed to identify and validate key nonverbal cues for assessing physician empathy in telemedicine consultations through a Delphi method. A three-round Delphi study was conducted from June to November 2022, involving various experts, including academics, healthcare professionals, AI/telemedicine researchers, industry professionals, and patients. Experts evaluated the importance, validity, and reliability of potential nonverbal cues. Consensus was determined based on median responses and expert scoring percentages, with statistical agreement and stability assessed using Kendall's coefficient of concordance (Kendall's W) and Wilcoxon signed-rank test. Analyses were conducted using SPSS, version 23.0 with significance set at p < 0.05. Of the 72 experts invited, 37 (51%) agreed to participate, with 35 completing the first round (95% completion rate). Eight significant nonverbal cues were identified in the first round, though one did not reach consensus. The second round obtained an 89% response rate (31/35), with three new cues introduced; one did not reach consensus. Round 3 achieved a 94% response rate (29/31), finalizing nine key cues: facial expression, eye contact, tone of voice, smiling, head nodding, body posture, hand gesture, distance, and environmental cues. Among these, facial expression, eye contact, and tone of voice were identified as the most crucial. Inter-expert agreement was statistically significant across all items with strong agreement on the importance (W = 0.739, p < 0.001), good agreement on their validity (W = 0.689, p < 0.001), and moderate agreement on their reliability (W = 0.452, p < 0.001). This study highlights the importance of specific nonverbal cues in telemedicine, particularly facial expression, eye contact, and tone of voice. It provides a validated foundation for developing tools to enhance physician-patient interactions and potentially improve health outcomes in telemedicine.
- 8 May 2025
- Published by Taylor & Francis in Medical Education Online
- Vol. 30  (1) , 2497328
Open AccessValidating nonverbal cues for assessing physician empathy in telemedicine: a Delphi study
Nonverbal communication is essential in physician-patient interaction, especially in telemedicine where verbal cues may be limited. This study aimed to identify and validate key nonverbal cues for assessing physician empathy in telemedicine consultations through a Delphi method. A three-round Delphi study was conducted from June to November 2022, involving various experts, including academics, healthcare professionals, AI/telemedicine researchers, industry professionals, and patients. Experts evaluated the importance, validity, and reliability of potential nonverbal cues. Consensus was determined based on median responses and expert scoring percentages, with statistical agreement and stability assessed using Kendall's coefficient of concordance (Kendall's W) and Wilcoxon signed-rank test. Analyses were conducted using SPSS, version 23.0 with significance set at p < 0.05. Of the 72 experts invited, 37 (51%) agreed to participate, with 35 completing the first round (95% completion rate). Eight significant nonverbal cues were identified in the first round, though one did not reach consensus. The second round obtained an 89% response rate (31/35), with three new cues introduced; one did not reach consensus. Round 3 achieved a 94% response rate (29/31), finalizing nine key cues: facial expression, eye contact, tone of voice, smiling, head nodding, body posture, hand gesture, distance, and environmental cues. Among these, facial expression, eye contact, and tone of voice were identified as the most crucial. Inter-expert agreement was statistically significant across all items with strong agreement on the importance (W = 0.739, p < 0.001), good agreement on their validity (W = 0.689, p < 0.001), and moderate agreement on their reliability (W = 0.452, p < 0.001). This study highlights the importance of specific nonverbal cues in telemedicine, particularly facial expression, eye contact, and tone of voice. It provides a validated foundation for developing tools to enhance physician-patient interactions and potentially improve health outcomes in telemedicine.