ABSTRACT
PURPOSE: Given the increasing significance and potential impact of artificial intelligence (AI) technology on health care delivery, there is an increasing demand to integrate AI into medical school curricula. This study aimed to define medical AI competencies and identify the essential competencies for medical graduates in South Korea. METHOD: An initial Delphi survey conducted in 2022 involving 4 groups of medical AI experts (n = 28) yielded 42 competency items. Subsequently, an online questionnaire survey was carried out with 1,955 participants (1,174 students and 781 professors) from medical schools across South Korea, utilizing the list of 42 competencies developed from the first Delphi round. A subsequent Delphi survey was conducted with 33 medical educators from 21 medical schools to differentiate the essential AI competencies from the optional ones. RESULTS: The study identified 6 domains encompassing 36 AI competencies essential for medical graduates: (1) understanding digital health and changes driven by AI; (2) fundamental knowledge and skills in medical AI; (3) ethics and legal aspects in the use of medical AI; (4) medical AI application in clinical practice; (5) processing, analyzing, and evaluating medical data; and (6) research and development of medical AI, as well as subcompetencies within each domain. While numerous competencies within the first 4 domains were deemed essential, a higher percentage of experts indicated responses in the last 2 domains, data science and medical AI research and development, were optional. CONCLUSIONS: This medical AI framework of 6 competencies and their subcompetencies for medical graduates exhibits promising potential for guiding the integration of AI into medical curricula. Further studies conducted in diverse contexts and countries are necessary to validate and confirm the applicability of these findings. Additional research is imperative for developing specific and feasible educational models to integrate these proposed competencies into pre-existing curricula.
Subject(s)
Artificial Intelligence , Curriculum , Delphi Technique , Schools, Medical , Students, Medical , Republic of Korea , Humans , Surveys and Questionnaires , Curriculum/standards , Schools, Medical/standards , Students, Medical/statistics & numerical data , Male , Female , Clinical Competence/standards , Adult , Faculty, MedicalABSTRACT
Fusion genes have been implicated in the development and progression of several types of sarcomas, serving as valuable diagnostic and prognostic markers, as well as potential therapeutic targets. We discovered a novel major facilitator superfamily domain-containing 7 (MFSD7) and adenosine triphosphate 5I (ATP5I) gene fusion from sarcomas. In this study, the MFSD7-ATP5I fusion transcript was screened using RNA sequencing in 55 sarcoma samples and sixteen normal samples. The MFSD7-ATP5I fusion transcript was detected in 58% of sarcoma samples. The correlation between the expression of MFSD7-ATP5I fusion transcript and clinicopathological information was analyzed, and MFSD7-ATP5I expression is associated with marked pleomorphism and lower tumor necrosis. Cell migration and invasion was significantly reduced by knock-down of MFSD7-ATP5I. Cell migration and invasion was increased by overexpression of MFSD7-ATP5I. A phosphokinase assay demonstrated that MFSD7-ATP5I is involved in the GSK-3 pathway. The current study found that MFSD7-ATP5I is associated with increasing pleomorphism and decreasing necrosis of tumors. And our gain and loss of function experiments prove that MFSD7-ATP5I promotes the invasiveness of tumor cells.
Subject(s)
Sarcoma , Soft Tissue Neoplasms , Humans , Glycogen Synthase Kinase 3 , Sarcoma/genetics , Cell Movement , NecrosisABSTRACT
BACKGROUND: Patients with relapsed osteosarcoma have poor treatment outcomes. High-dose chemotherapy with autologous stem cell transplantation (HDCT/ASCT) has been used in several high-risk malignant solid tumors; however, few studies have evaluated their role in treating osteosarcoma. We evaluated the effectiveness of HDCT/ASCT in relapsed pediatric osteosarcoma cases. PROCEDURE: We retrospectively reviewed the medical records of 40 patients diagnosed with and treated for relapsed osteosarcoma at Asan Medical Center and Samsung Medical Center from January 1996 to July 2019. RESULTS: The median age of this cohort was 13.4 years (range: 6.1-18.2). The cohort's 5-year overall survival (OS) was 51.0% ± 0.1% during a median follow-up period of 67.5 months. Twenty-five patients (62.5%) achieved complete remission (CR) with salvage treatment, and the 5-year OS was 82.4% ± 0.1%, whereas none of the remaining 15 patients who did not achieve CR survived (p < .0001). Of the 25 CR cases, 15 underwent subsequent HDCT/ASCT. We compared the effect of HDCT/ASCT among patients who achieved CR. There were no significant differences in the 5-year OS outcomes between patients who did and did not receive HDCT/ASCT (83.9% ± 0.1%, 13/15 vs. 80.0% ± 0.1%, 8/10, respectively; p = .923). CONCLUSION: To our knowledge, we report the first comparative cohort study that proved HDCT/ASCT does not significantly improve survival outcomes in relapsed osteosarcoma. Achievement of CR remains the most crucial factor for good survival outcomes.
Subject(s)
Hematopoietic Stem Cell Transplantation , Osteosarcoma , Humans , Child , Adolescent , Retrospective Studies , Cohort Studies , Antineoplastic Combined Chemotherapy Protocols/therapeutic use , Transplantation, Autologous , Disease-Free Survival , Stem Cell TransplantationABSTRACT
BACKGROUND: The recent explosion of cancer genomics provides extensive information about mutations and gene expression changes in cancer. However, most of the identified gene mutations are not clinically utilized. It remains uncertain whether the presence of a certain genetic alteration will affect treatment response. Conventional statistics have limitations for causal inferences and are hard to gain sufficient power in genomic datasets. Here, we developed and evaluated a C-search algorithm for searching the causal genes that maximize the effect of the treatment. METHODS: The algorithm was developed based on the potential outcome framework and Bayesian posterior update. The precision of the algorithm was validated using a simulation dataset. The algorithm was implemented to a cBioPortal dataset. The genes discovered by the algorithm were externally validated within CancerSCAN screening data from Samsung Medical Center. RESULTS: Simulation data analysis showed that the C-search algorithm was able to identify nine causal genes out of ten. The C-search algorithm shows the discovery rate rapidly increasing until the 1500 data instances. Meanwhile, the log-rank test shows a slower increase in performance. The C-search algorithm was able to suggest nine causal genes from the cBioPortal Metabric dataset. Treating the patients with the causal genes is associated with better survival outcome in both the cBioPortal dataset and the CancerSCAN dataset which is used for external validation. CONCLUSIONS: Our C-search algorithm demonstrated better performance to identify causal effects of the genes than multiple log-rank test analysis especially within a limited number of data. The result suggests that the C-search can discover the causal genes from various genetic datasets, where the number of samples is limited compared to the number of variables.
Subject(s)
Algorithms , Genomics , Bayes Theorem , Causality , Data Collection , HumansABSTRACT
PURPOSE: Early detection and classification of bone tumors in the proximal femur is crucial for their successful treatment. This study aimed to develop an artificial intelligence (AI) model to classify bone tumors in the proximal femur on plain radiographs. METHODS: Standard anteroposterior hip radiographs were obtained from a single tertiary referral center. A total of 538 femoral images were set for the AI model training, including 94 with malignant, 120 with benign, and 324 without tumors. The image data were pre-processed to be optimized for training of the deep learning model. The state-of-the-art convolutional neural network (CNN) algorithms were applied to pre-processed images to perform three-label classification (benign, malignant, or no tumor) on each femur. The performance of the CNN model was verified using fivefold cross-validation and was compared against that of four human doctors. RESULTS: The area under the receiver operating characteristic (AUROC) of the best performing CNN model for the three-label classification was 0.953 (95% confidence interval, 0.926-0.980). The diagnostic accuracy of the model (0.853) was significantly higher than that of the four doctors (0.794) (P = 0.001) and also that of each doctor individually (0.811, 0.796, 0.757, and 0.814, respectively) (P<0.05). The mean sensitivity, specificity, precision, and F1 score of the CNN models were 0.822, 0.912, 0.829, and 0.822, respectively, whereas the mean values of four doctors were 0.751, 0.889, 0.762, and 0.797, respectively. CONCLUSIONS: The AI-based model demonstrated high performance in classifying the presence of bone tumors in the proximal femur on plain radiographs. Our findings suggest that AI-based technology can potentially reduce the misdiagnosis of doctors who are not specialists in musculoskeletal oncology.
Subject(s)
Artificial Intelligence , Bone Neoplasms/classification , Femur , Radiography/methods , Algorithms , Bone Neoplasms/diagnostic imaging , Humans , Image Processing, Computer-Assisted , Neural Networks, Computer , Observer Variation , ROC Curve , Reproducibility of ResultsABSTRACT
Disruption of bone homeostasis caused by metastatic osteolytic breast cancer cells increases inflammatory osteolysis and decreases bone formation, thereby predisposing patients to pathological fracture and cancer growth. Alteration of osteoblast function induces skeletal diseases due to the disruption of bone homeostasis. We observed increased activation of pERK1/2 in osteolytic breast cancer cells and osteoblasts in human pathological specimens with aggressive osteolytic breast cancer metastases. We confirmed that osteolytic breast cancers with high expression of pERK1/2 disrupt bone homeostasis via osteoblastic ERK1/2 activation at the bone-breast cancer interface. The process of inflammatory osteolysis modulates ERK1/2 activation in osteoblasts and breast cancer cells through dominant-negative MEK1 expression and constitutively active MEK1 expression to promote cancer growth within bone. Trametinib, an FDA-approved MEK inhibitor, not only reduced breast cancer-induced bone destruction but also dramatically reduced cancer growth in bone by inhibiting the inflammatory skeletal microenvironment. Taken together, these findings suggest that ERK1/2 activation in both breast cancer cells and osteoblasts is required for osteolytic breast cancer-induced inflammatory osteolysis and that ERK1/2 pathway inhibitors may represent a promising adjuvant therapy for patients with aggressive osteolytic breast cancers by altering the shared cancer and bone microenvironment.
ABSTRACT
Liposarcoma (LPS) is an adult soft tissue malignancy that arises from fat tissue, where well-differentiated (WD) and dedifferentiated (DD) forms are the most common. DDLPS represents the progression of WDLPS into a more aggressive high-grade and metastatic form. Although a few DNA copy-number amplifications are known to be specifically found in WD- or DDLPS, systematic genetic differences that signify subtype determination between WDLPS and DDLPS remain unclear. Here, we profiled the genome and transcriptome of 38 LPS tumors to uncover the genetic signatures of subtype differences. Replication-dependent histone (RD-HIST) mRNAs were highly elevated and their regulation was disrupted in a subset of DDLPS, increasing cellular histone molecule levels, as measured using RNA-seq (the averaged fold change of 53 RD-HIST genes between the DD and WD samples was 10.9) and immunohistochemistry. The change was not observed in normal tissues. Integrated whole-exome sequencing, RNA-seq, and methylation analyses revealed that the overexpressed HMGA2 (the fold change between DD and WD samples was 7.3) was responsible for the increased RD-HIST level, leading to aberrant cell proliferation. Therefore, HMGA2-mediated elevation of RD-HISTs were crucial events in determining the aggressiveness of DDLPS, which may serve as a biomarker for prognosis prediction for liposarcoma patients.
ABSTRACT
AIMS: Bone metastasis ultimately occurs due to a complex multistep process, during which the interactions between cancer cells and bone microenvironment play important roles. Prior to colonization of the bone, cancer cells must succeed through a series of steps that will allow them to gain migratory and invasive properties; epithelial-to-mesenchymal transition (EMT) is known to be integral here. The aim of this study was to determine the effects of G protein subunit alpha Q (GNAQ) on the mechanisms underlying bone metastasis through EMT pathway. METHODS: A total of 80 tissue samples from patients who were surgically treated during January 2012 to December 2014 were used in the present study. Comparative gene analysis revealed that the GNAQ was more frequently altered in metastatic bone lesions than in primary tumour sites in lung cancer patients. We investigated the effects of GNAQ on cell proliferation, migration, EMT, and stem cell transformation using lung cancer cells with GNAQ-knockdown. A xenograft mouse model tested the effect of GNAQ using micro-CT analyses and histological analyses. RESULTS: GNAQ-knockdown showed down-regulation of tumour growth through mitogen-activated protein kinase (MAPK) signalling in lung cancer cells, but not increased apoptosis. We found that GNAQ-knockdown induced EMT and promoted invasiveness. GNAQ-knockdown cells injected into the bone marrow of murine tibia induced tumour growth and bone-to-lung metastasis, whereas it did not in control mice. Moreover, the knockdown of GNAQ enhanced cancer stem cell-like properties in lung cancer cells, which resulted in the development of resistance to chemotherapy. CONCLUSION: The present study reveals that the GNAQ-knockdown induced cancer stem cell-like properties. Cite this article: Bone Joint Res 2021;10(5):310-320.
ABSTRACT
Acquisition of metastatic potential by cancer cells is related to cancer stemness and anchorage-independent growth. The onset and progression of cancer are known to involve Hedgehog (HH) signaling that is activated by the binding of HH to the Patched 1 (PTCH1) receptor. However, the functions and mechanisms of action of PTCH1 in the context of bone metastasis remain to be elucidated. In this study, lentivirally-delivered shRNA was used to deplete PTCH1 levels, which resulted in the inhibition of spherical colony formation by the human non-small cell lung cancer (NSCLC) cell line; this suggested that PTCH1 promotes anchorage-independent growth. Concordantly, knockdown of PTCH1 resulted in significantly reduced migration and invasion of NSCLC cells; this was accompanied by the downregulation of MMP7 and SOX2. PTCH1 knockdown resulted in decreased bone destruction and osteoclastogenesis in a mouse bone metastasis model. These results indicate that PTCH1 may be an important regulator of bone invasion, and strongly suggest that knockdown of PTCH1 may decrease the anchorage-independent growth and metastatic potential of NSCLC.
Subject(s)
Carcinoma, Non-Small-Cell Lung , Lung Neoplasms , Patched-1 Receptor , Animals , Bone Neoplasms/secondary , Carcinoma, Non-Small-Cell Lung/genetics , Carcinoma, Non-Small-Cell Lung/pathology , Cell Line, Tumor , Hedgehog Proteins , Humans , Lung Neoplasms/genetics , Lung Neoplasms/pathology , Mice , Patched-1 Receptor/genetics , Patched-1 Receptor/metabolism , Signal TransductionABSTRACT
BACKGROUND: The impact of adjuvant chemotherapy or radiation therapy on the survival of patients with synovial sarcoma (SS), which is a rare soft-tissue sarcoma, remains controversial. Bayesian statistical approaches and propensity score matching can be employed to infer treatment effects using observational data. Thus, this study aimed to identify the individual treatment effects of adjuvant therapies on the overall survival of SS patients and recognize subgroups of patients who can benefit from specific treatments using Bayesian subgroup analyses. METHODS: We analyzed data from patients with SS obtained from the surveillance, epidemiology, and end results (SEER) public database. These data were collected between 1984 and 2014. The treatment effects of chemotherapy and radiation therapy on overall survival were evaluated using propensity score matching. Subgroups that could benefit from radiation therapy or chemotherapy were identified using Bayesian subgroup analyses. RESULTS: Based on a stratified Kaplan-Meier curve, chemotherapy exhibited a positive average causal effect on survival in patients with SS, whereas radiation therapy did not. The optimal subgroup for chemotherapy includes the following covariates: older than 20 years, male, large tumor (longest diameter > 5 cm), advanced stage (SEER 3), extremity location, and spindle cell type. The optimal subgroup for radiation therapy includes the following covariates: older than 20 years, male, large tumor (longest diameter > 5 cm), early stage (SEER 1), extremity location, and biphasic type. CONCLUSION: In this study, we identified high-risk patients whose variables include age (age > 20 years), gender, tumor size, tumor location, and poor prognosis without adjuvant treatment. Radiation therapy should be considered in the early stages for high-risk patients with biphasic types. Conversely, chemotherapy should be considered for late-stage high-risk SS patients with spindle cell types.
Subject(s)
Chemotherapy, Adjuvant/methods , Radiotherapy/methods , Sarcoma, Synovial/therapy , Bayes Theorem , Combined Modality Therapy , Female , Humans , Kaplan-Meier Estimate , Male , Retrospective Studies , Sarcoma, Synovial/mortality , Sarcoma, Synovial/pathology , Survival Rate , Treatment OutcomeABSTRACT
In recent years, artificial intelligence (AI) technologies have greatly advanced and become a reality in many areas of our daily lives. In the health care field, numerous efforts are being made to implement the AI technology for practical medical treatments. With the rapid developments in machine learning algorithms and improvements in hardware performances, the AI technology is expected to play an important role in effectively analyzing and utilizing extensive amounts of health and medical data. However, the AI technology has various unique characteristics that are different from the existing health care technologies. Subsequently, there are a number of areas that need to be supplemented within the current health care system for the AI to be utilized more effectively and frequently in health care. In addition, the number of medical practitioners and public that accept AI in the health care is still low; moreover, there are various concerns regarding the safety and reliability of AI technology implementations. Therefore, this paper aims to introduce the current research and application status of AI technology in health care and discuss the issues that need to be resolved.
Subject(s)
Artificial Intelligence , Delivery of Health Care , Government Regulation , Health Policy , Humans , Image Processing, Computer-Assisted , Magnetic Resonance Imaging , Safety Management , Tomography, X-Ray ComputedABSTRACT
AIMS: Receptor activator of nuclear factor-κB ligand (RANKL) is a key molecule that is expressed in bone stromal cells and is associated with metastasis and poor prognosis in many cancers. However, cancer cells that directly express RANKL have yet to be unveiled. The current study sought to evaluate how a single subunit of G protein, guanine nucleotide-binding protein G(q) subunit alpha (GNAQ), transforms cancer cells into RANKL-expressing cancer cells. METHODS: We investigated the specific role of GNAQ using GNAQ wild-type cell lines (non-small-cell lung cancer cell lines; A549 cell lines), GNAQ knockdown cell lines, and patient-derived cancer cells. We evaluated GNAQ, RANKL, macrophage colony-stimulating factor (M-CSF), nuclear transcription factor-κB (NF-κB), inhibitor of NF-κB (IκB), and protein kinase B (Akt) signalling in the GNAQ wild-type and the GNAQ-knockdown cells. Osteoclastogenesis was also evaluated in both cell lines. RESULTS: In the GNAQ-knockdown cells, RANKL expression was significantly upregulated (p < 0.001). The expression levels of M-CSF were also significantly increased in the GNAQ-knockdown cells compared with control cells (p < 0.001). GNAQ knockdown cells were highly sensitive to tumour necrosis factor alpha (TNF-α) and showed significant activation of the NF-κB pathway. The expression levels of RANKL were markedly increased in GNAQ mutant compared with GNAQ wild-type in patient-derived tumour tissues. CONCLUSION: The present study reveals that the alterations of GNAQ activate NF-κB pathway in cancers, which increase RANKL and M-CSF expression and induce osteoclastogenesis in cancers.Cite this article: Bone Joint Res. 2020;9(1):29-35.
ABSTRACT
OBJECTIVES: To compare magnetic resonance imaging (MRI) parameters of small, deeply located non-malignant and malignant soft-tissue tumors (STTs). METHODS: Between May 2011 and December 2017, 95 MRIs in 95 patients with pathologically proven STTs of small size (<5 cm) and deep location (66 non-malignant and 29 malignant) were identified. For qualitative parameters, consensus reading was performed by three radiologists for presence of necrosis, infiltration, lobulation, and the tail sign. Apparent diffusion coefficient (ADC) was analyzed by two other radiologists independently. Univariable and multivariable analyses were performed to determine the diagnostic performances of MRI parameters in differentiating non-malignancy and malignancy, and for non-myxoid, non-hemosiderin STTs and myxoid STTs as subgroups. Interobserver agreement for ADC measurement was calculated with the intraclass correlation coefficient. RESULTS: Interobserver agreement on ADC measurement was almost perfect. On univariable analysis, the malignant group showed a significantly larger size, lower ADC, and higher incidence of all qualitative MRI parameters for all STTs. Size (p = 0.012, odds ratio [OR] 2.57), ADC (p = 0.041, OR 3.85), and the tail sign (p = 0.009, OR 6.47) were independently significant on multivariable analysis. For non-myxoid, non-hemosiderin STTs, age, size, ADC, frequency of infiltration, lobulation, and the tail sign showed significant differences between non-malignancy and malignancy on univariable analysis. Only ADC (p = 0.032, OR 142.86) retained its independence on multivariable analysis. For myxoid STTs, only size and tail sign were significant on univariable analysis without independent significance. CONCLUSIONS: Size, ADC, and incidence of qualitative MRI parameters were significantly different between small, deeply located non-malignant and malignant STTs. Only ADC was independently significant for both overall analysis and the non-myxoid, non-hemosiderin subgroup.
Subject(s)
Soft Tissue Neoplasms/diagnosis , Soft Tissue Neoplasms/pathology , Adolescent , Adult , Aged , Aged, 80 and over , Child , Female , Humans , Image Interpretation, Computer-Assisted/methods , Magnetic Resonance Imaging/methods , Male , Middle Aged , Retrospective Studies , Young AdultABSTRACT
OBJECTIVE: In this study, we established a risk scoring system using easily obtained clinical characteristics at the time of initiating palliative chemotherapy to predict accurate overall survival of patients with advanced gastric cancer after first-line treatment with fluoropyrimidine-platinum combination chemotherapy. METHODS: A total of 1733 patients treated at the Samsung Medical Center, Korea were included in the study, and clinicopathological and laboratory data were retrospectively analysed. The dataset was split into a training set (n=1156, 67%) and a validation set (n=577, 33%). Top-ranked variables were identified using the random forest survival algorithm and integrated into a Cox regression model, thereby constructing the scoring system for predicting the overall survival of patients with advanced gastric cancer. RESULTS: The following five variables were finally included in the scoring system: serum neutrophil-lymphocyte ratio, alkaline phosphatase level, albumin level, performance status and histologic differentiation. The scoring system determined four distinct risk groups in the validation dataset with median overall survival of 17.1 months (95% CI=14.9 to 20.5 months), 12.9 months (95% CI=11.4 to 14.6 months), 8.1 months (95% CI=5.3 to 12.3 months) and 3.9 months (95% CI=1.5 to 8.2 months), respectively. The area under the curve to estimate the discrimination performance of the scoring system was 66.1 considering 1 year overall survival. CONCLUSIONS: We developed a simple and clinically useful predictive scoring model in a homogeneous population with advanced gastric cancer treated with fluoropyrimidine-containing and platinum-containing chemotherapy. However, additional independent validation will be required before the scoring model can be used commonly.
Subject(s)
Stomach Neoplasms/mortality , Female , Humans , Male , Middle Aged , PrognosisABSTRACT
BACKGROUND: We used the Surveillance, Epidemiology, and End Results (SEER) database to develop and validate deep survival neural network machine learning (ML) algorithms to predict survival following a spino-pelvic chondrosarcoma diagnosis. METHODS: The SEER 18 registries were used to apply the Risk Estimate Distance Survival Neural Network (RED_SNN) in the model. Our model was evaluated at each time window with receiver operating characteristic curves and areas under the curves (AUCs), as was the concordance index (c-index). RESULTS: The subjects (n = 1088) were separated into training (80%, n = 870) and test sets (20%, n = 218). The training data were randomly sorted into training and validation sets using 5-fold cross validation. The median c-index of the five validation sets was 0.84 (95% confidence interval 0.79-0.87). The median AUC of the five validation subsets was 0.84. This model was evaluated with the previously separated test set. The c-index was 0.82 and the mean AUC of the 30 different time windows was 0.85 (standard deviation 0.02). According to the estimated survival probability (by 62 months), we divided the test group into five subgroups. The survival curves of the subgroups showed statistically significant separation (p < 0.001). CONCLUSIONS: This study is the first to analyze population-level data using artificial neural network ML algorithms for the role and outcomes of surgical resection and radiation therapy in spino-pelvic chondrosarcoma.
Subject(s)
Bone Neoplasms , Chondrosarcoma , Humans , Machine Learning , Neural Networks, Computer , PatientsABSTRACT
OBJECTIVES: To examine the correlation of diffusion-weighted and dynamic contrast-enhanced magnetic resonance imaging (MRI) parameters with Ki-67 labeling index (LI) in soft tissue sarcoma (STS). METHODS: The institutional review board approved this retrospective study, and the requirement for informed consent was waived. Thirty-six patients with STS who underwent 3.0-T MRI, including diffusion-weighted and dynamic contrast-enhanced MRI, between July 2011 and February 2018, were included in this study. The mean and minimum apparent diffusion coefficients (ADCs) (ADCmean and ADCmin, respectively), volume transfer constant, reflux rate, and volume fraction of the extravascular extracellular matrix of each lesion were independently analyzed by two readers. Their relationship with the Ki-67 LI was examined using Spearman's correlation analyses. Differences between low- and high-proliferation groups based on Ki-67 LI were evaluated statistically. Optimal cut-off points were determined using the area under the curve analysis for significant parameters. Interobserver agreement was assessed with the intraclass correlation coefficient. RESULTS: ADCmean (ρ = - 0.333, p = 0.047) was significantly and inversely correlated with Ki-67 LI. The high-proliferation group showed a significantly lower ADCmean than did the low-proliferation group (median, 1.08 vs. 1.20; p = 0.048). When a cut-off ADCmean value of 1.16 × 10-3 mm2/s was used, the sensitivity, specificity, and area under the curve for differentiating low- and high-proliferation groups were 75.0%, 60.0%, and 0.712, respectively. Interobserver agreements between the two readers were almost perfect for all parameters. CONCLUSIONS: ADCmean was correlated with Ki-67 LI and could help differentiate between STS with low and high proliferation potential. KEY POINTS: ⢠ADC meanwas significantly and inversely correlated with Ki-67 labeling index in soft tissue sarcoma. ⢠In the high-proliferation group, ADC meanvalues were significantly lower than those of the low-proliferation group.
Subject(s)
Sarcoma/pathology , Soft Tissue Neoplasms/pathology , Adult , Aged , Aged, 80 and over , Biomarkers, Tumor/metabolism , Cell Proliferation/physiology , Contrast Media , Diffusion Magnetic Resonance Imaging/methods , Female , Humans , Ki-67 Antigen/metabolism , Magnetic Resonance Imaging/methods , Male , Middle Aged , Retrospective Studies , Sensitivity and SpecificityABSTRACT
OBJECTIVES: The aim of this study was to develop a new prognostic classification for epithelial ovarian cancer (EOC) patients using gradient boosting (GB) and to compare the accuracy of the prognostic model with the conventional statistical method. METHODS: Information of EOC patients from Samsung Medical Center (training cohort, n=1,128) was analyzed to optimize the prognostic model using GB. The performance of the final model was externally validated with patient information from Asan Medical Center (validation cohort, n=229). The area under the curve (AUC) by the GB model was compared to that of the conventional Cox proportional hazard regression analysis (CoxPHR) model. RESULTS: In the training cohort, the AUC of the GB model for predicting second year overall survival (OS), with the highest target value, was 0.830 (95% confidence interval [CI]=0.802-0.853). In the validation cohort, the GB model also showed high AUC of 0.843 (95% CI=0.833-0.853). In comparison, the conventional CoxPHR method showed lower AUC (0.668 (95% CI=0.617-0.719) for the training cohort and 0.597 (95% CI=0.474-0.719) for the validation cohort) compared to GB. New classification according to survival probability scores of the GB model identified four distinct prognostic subgroups that showed more discriminately classified prediction than the International Federation of Gynecology and Obstetrics staging system. CONCLUSION: Our novel GB-guided classification accurately identified the prognostic subgroups of patients with EOC and showed higher accuracy than the conventional method. This approach would be useful for accurate estimation of individual outcomes of EOC patients.