• Journal of Gastric Cancer
• /
• v.22 no.2
• /
• pp.120-134
• /
• 2022

Purpose: This study aimed to identify prognostic factors for patients with distant lymph node-involved gastric cancer (GC) using a machine learning algorithm, a method that offers considerable advantages and new prospects for high-dimensional biomedical data exploration. Materials and Methods: This study employed 79 features of clinical pathology, laboratory tests, and therapeutic details from 289 GC patients whose distant lymphadenopathy was presented as the first episode of recurrence or metastasis. Outcomes were measured as any-cause death events and survival months after distant lymph node metastasis. A prediction model was built based on possible outcome predictors using a random survival forest algorithm and confirmed by 5×5 nested cross-validation. The effects of single variables were interpreted using partial dependence plots. A contour plot was used to visually represent survival prediction based on 2 predictive features. Results: The median survival time of patients with GC with distant nodal metastasis was 9.2 months. The optimal model incorporated the prealbumin level and the prothrombin time (PT), and yielded a prediction error of 0.353. The inclusion of other variables resulted in poorer model performance. Patients with higher serum prealbumin levels or shorter PTs had a significantly better prognosis. The predicted one-year survival rate was stratified and illustrated as a contour plot based on the combined effect the prealbumin level and the PT. Conclusions: Machine learning is useful for identifying the important determinants of cancer survival using high-dimensional datasets. The prealbumin level and the PT on distant lymph node metastasis are the 2 most crucial factors in predicting the subsequent survival time of advanced GC.

• Journal of the Korean Geotechnical Society
• /
• v.40 no.3
• /
• pp.77-84
• /
• 2024

Aging and damaged underground utilities cause cavity and ground subsidence under roads, which can cause economic losses and risk user safety. This study used infrared cameras to assess the thermal characteristics of such cavities and evaluate their reliability using a CNN algorithm. PVC pipes were embedded at various depths in a test site measuring 400 cm × 50 cm × 40 cm. Concrete blocks were used to simulate road surfaces, and measurements were taken from 4 PM to noon the following day. The initial temperatures measured by the infrared camera were 43.7℃, 43.8℃, and 41.9℃, reflecting atmospheric temperature changes during the measurement period. The RP algorithm generates images in four resolutions, i.e., 10,000 × 10,000, 2,000 × 2,000, 1,000 × 1,000, and 100 × 100 pixels. The accuracy of the CNN model using RP images as input was 99%, 97%, 98%, and 96%, respectively. These results represent a considerable improvement over the 73% accuracy obtained using time-series images, with an improvement greater than 20% when using the RP algorithm-based inputs.

• Journal of the Korean Orthopaedic Association
• /
• v.54 no.1
• /
• pp.37-44
• /
• 2019

Purpose: Tumor infiltration around the knee joint or skip metastasis, repeated infection sequelae after tumor prosthesis implantation, regional recurrence, and mechanical failure of the megaprosthesis might require combined distal femur and proximal tibia replacement (CFTR). Among the aforementioned situations, there are few reports on the indication, complications, and implant survival of CFTR in temporarily arthrodesed patients who had a massive bony defect on either side of the knee joint to control infection. Materials and Methods: Thirty-four CFTR patients were reviewed retrospectively and 13 temporary arthrodesed cases switched to CFTR were extracted. All 13 cases had undergone a massive bony resection on either side of the knee joint and temporary arthrodesis state to control the repeated infection. This paper describes the diagnosis, tumor location, number of operations until CFTR, duration from the index operation to CFTR, survival of CFTR, complications, and Musculoskeletal Tumor Society (MSTS) score. Results: According to Kaplan-Meier plot, the 5- and 10-year survival of CFTR was 69.0%±12.8%, 46.0%±20.7%, respectively. Six (46.2%) of the 13 cases had major complications. Three cases underwent removal of the prosthesis and were converted to arthrodesis due to infection. Two cases underwent partial change of the implant due to loosening and periprosthetic fracture. The remaining case with a deep infection was resolved after extensive debridement. At the final follow-up, the average MSTS score of 10 cases with CFTR was 24.6 (21-27). In contrast, the MSTS score of 3 arthrodesis cases with failed CFTR was 12.3 (12-13). The average range of motion of the 10 CFTR cases was 67° (0°-100°). The mean extension lag of 10 cases was 48° (20°-80°). Conclusion: Although the complication rates is substantial, conversion of an arthrodesed knee to a mobile joint using CFTR in a patient who had a massive bony defect on either side of the knee joint to control infection should be considered. The patient's functional outcome was different from the arthrodesed one. For successful conversion to a mobile joint, thorough the eradication of scar tissue and creating sufficient space for the tumor prosthesis to flex the knee joint up to 60° to 70° without soft tissue tension.