Speakers

Outcomes in paediatric traumatic brain injury What AI has to offer
1108 14:20-14:30
Neuro-trauma & Intensive Care/303A
BACKGROUND- Healthcare artificial a general of machine learning perception in analyzing, presenting, and understanding healthcare data.goal of this study predict METHODOLOGY- The present study using machine learning approach was carried out in Department of Neurosurgery SMS, Medical college and Hospital, Jaipur Rajasthan from January 2020 to December 2022. At the time of admission GCS score was calculated, CT scan was performed and other investigations were done. We utilized the PYTHON Network Toolbox to construct an ANN that classified patients into two groups, either “favourable 6- month outcomes” or “unfavourable 6-month outcomes,” RESULTS- 54% of patients who had unfavourable outcome midline shift was found to be >5mm while only 3% patients who had favourable outcome had midline shift > 5mm. Of all the patients who had unfavourable outcome 82% patient had no eye opening and 64% patients who had favourable outcome showed eye opening to speech or had spontaneous eye opening. All the patients who had unfavourable outcome 79% patient had no verbal response .Only 14% of patients who showed favourable outcome had verbal response ≤V2. All the patients who had unfavourable outcome,43% patient had motor response of ≤M2 and only 4% of patients who showed favourable outcome had motor response ≤M3.The deep learning model was validated on the validation set and the AUC score was found to be 0.99 with an accuracy of 0.937. CONCLUSIONS -This study showed that machine-learning can be leveraged to more accurately predict TBI outcomes in children.

3D Technology in neurosurgery
1109 08:55-09:05
AI & New Technology/304B
Background: Recent advancements in three‐dimensional (3D) printing technology in the field of neurosurgery have given a newer modality of management for patients. In this article, we intend to share our institutional experience regarding the use of 3D printing in three modalities, namely, cranioplasty using customized 3D‐printed molds of polymethylmethacrylate, 3D‐printed model‐assisted management of craniovertebral (CV) junction abnormalities, and 3D model‐assisted management of brain tumors. Materials and Methods: A total of 55 patients were included in our study at S. M. S Medical College, Jaipur, India. 3D‐printed models were prepared for cranioplasty in 30 cases, CV junction anomalies in 18 cases, and brain tumors in 7 cases. Preoperative and postoperative data were analyzed as per the diagnosis. Results: In cranioplasty, cranial contour and approximation of the margins were excellent and esthetic appearance improved in all patients. In CV junction anomalies, neck pain and myelopathy were improved in all patients, as analyzed using the visual analog scale and the Japanese Orthopedic Association Scale score, respectively. Our questionnaire survey revealed that 3D models for brain tumors were useful in understanding space interval and depth intraoperatively with added advantage of patient education. Conclusion: Rapid prototyping 3D‐printing technologies provide a practical and anatomically accurate means to produce patient‐specific and disease‐specific models. These models allow for surgical planning, training, simulation, and devices for the assessment and treatment of neurosurgical disease. Expansion of this technology in neurosurgery will serve practitioners, trainees, and patients.