Enhancing gestational diabetes mellitus risk assessment and treatment through GDMPredictor: a machine learning approach.

BACKGROUND: Gestational diabetes mellitus (GDM) is a serious health concern that affects pregnant women worldwide and can lead to adverse pregnancy outcomes. Early detection of high-risk individuals and the implementation of appropriate treatment can enhance these outcomes. METHODS: We conducted a study on a cohort of 3467 pregnant women during their pregnancy, with a total of 5649 clinical and biochemical records collected. We utilized this dataset as our training dataset to develop a web server called GDMPredictor. The GDMPredictor utilizes advanced machine learning techniques to predict the risk of GDM in pregnant women. We also personalize treatment recommendations based on essential biochemical indicators, such as A1MG, BMG, CysC, CO2, TBA, FPG, and CREA. Our assessment of GDMPredictor's effectiveness involved training it on the dataset of 3467 pregnant women and measuring its ability to predict GDM risk using an AUC and auPRC. RESULTS: GDMPredictor demonstrated an impressive level of precision by achieving an AUC score of 0.967. To tailor our treatment recommendations, we use the GDM risk level to identify higher risk candidates who require more intensive care. The GDMPredictor can accept biochemical indicators for predicting the risk of GDM at any period from 1 to 24 weeks, providing healthcare professionals with an intuitive interface to identify high-risk patients and give optimal treatment recommendations. CONCLUSIONS: The GDMPredictor presents a valuable asset for clinical practice, with the potential to change the management of GDM in pregnant women. Its high accuracy and efficiency make it a reliable tool for doctors to improve patient outcomes. Early identification of high-risk individuals and tailored treatment can improve maternal and fetal health outcomes http://www.bioinfogenetics.info/GDM/ .

Adversarial reinforcement learning for dynamic treatment regimes.

Treatment recommendation, as a critical task of delivering effective interventions according to patient state and expected outcome, plays a vital role in precision medicine and healthcare management. As a well-suited tactic to learn optimal policies of recommender systems, reinforcement learning is promising to address the challenge of treatment recommendation. However, existing solutions mostly require frequent interactions between treatment recommender systems and clinical environment, which are expensive, time-consuming, and even infeasible in clinical practice. In this study, we present a novel model-based offline reinforcement learning approach to optimize a treatment policy by utilizing patient treatment trajectories in Electronic Health Records (EHRs). Specifically, a patient treatment trajectory simulator is firstly constructed based on the ground-truth trajectories in EHRs. Thereafter, the constructed simulator is utilized to model the online interactions between the treatment recommender system and clinical environment. In this way, the counterfactual trajectories can be generated. To alleviate the bias deriving from the ground-truth and the counterfactual trajectories, an adversarial network is incorporated into the proposed model, such that a large space of treatment actions can be explored with the scaled rewards. The proposed model is evaluated on a simulated dataset and a real-world dataset. The experimental results demonstrate that the proposed model is superior to other methods, giving rise to a new solution for dynamic treatment regimes and beyond.

Swiss GPs' preferences for antidepressant treatment in mild depression: vignette-based quantitative analysis.

BACKGROUND: GPs frequently prescribe antidepressants in mild depression. The aim of this study was to examine, how often Swiss GPs recommend antidepressants in various clinical presentations of mild depression and which factors contribute to antidepressant treatment recommendations. METHODS: We conducted an online survey among Swiss GPs with within-subject effect analysis. Alternating case vignettes described a typical female case of mild depression according to International Classification of Diseases, 10th edition criteria, with and without anxiety symptoms and sleep problems. GPs indicated for each vignette their preferred treatments (several recommendations were possible). Additionally, we assessed GP characteristics, attitudes towards depression treatments, and elements of clinical decision-making. RESULTS: Altogether 178 GPs completed the survey. In the initial description of a case with mild depression, 11% (95%-CI: 7%-17%) of GPs recommended antidepressants. If anxiety symptoms were added to the same case, 29% (23%-36%) recommended antidepressants. If sleep problems were mentioned, 47% (40%-55%) recommended antidepressants, and if both sleep problems and anxiety symptoms were mentioned, 63% (56%-70%) recommended antidepressants. Several factors were independently associated with increased odds of recommending antidepressants, specifically more years of practical experience, an advanced training in psychosomatic and psychosocial medicine, self-dispensation, and a higher perceived effectiveness of antidepressants. By contrast, a higher perceived influence of patient characteristics and the use of clinical practice guidelines were associated with reduced odds of recommending antidepressants. CONCLUSIONS: Consistent with depression practice guidelines, Swiss GPs rarely recommended antidepressants in mild depression if no co-indications (i.e., sleep problems and anxiety symptoms) were depicted. However, presence of sleep problems and anxiety symptoms, many years of practical experience, overestimation of antidepressants' effectiveness, self-dispensation, an advanced training in psychosomatic and psychosocial medicine, and non-use of clinical practice guidelines may independently lead to antidepressant over-prescribing.

[Multiple sclerosis treatment consensus group (MSTCG): position paper on disease-modifying treatment of multiple sclerosis 2021 (white paper)]. / Multiple Sklerose Therapie Konsensus Gruppe (MSTKG): Positionspapier zur verlaufsmodifizierenden Therapie der Multiplen Sklerose 2021 (White Paper).

Multiple sclerosis is a complex, autoimmune-mediated disease of the central nervous system characterized by inflammatory demyelination and axonal/neuronal damage. The approval of various disease-modifying therapies and our increased understanding of disease mechanisms and evolution in recent years have significantly changed the prognosis and course of the disease. This update of the Multiple Sclerosis Therapy Consensus Group treatment recommendation focuses on the most important recommendations for disease-modifying therapies of multiple sclerosis in 2021. Our recommendations are based on current scientific evidence and apply to those medications approved in wide parts of Europe, particularly German-speaking countries (Germany, Austria, Switzerland).

Concordance between treatment recommendations provided by IBM Watson for Oncology and a multidisciplinary tumor board for breast cancer in China.

OBJECTIVE: Watson for Oncology (WFO), an artificial intelligence from IBM Corporation, can provide a treatment plan by analyzing patient's disease characteristics. The present study was performed to examine the concordance between treatment recommendations proposed by WFO and the multidisciplinary tumor board at our center. The aim was to explore the feasibility of using WFO for breast cancer cases in China and to ascertain the ways to make WFO more suitable for Chinese patients with breast cancer. METHODS: Data from 302 breast cancer patients treated at the Second Affiliated Hospital of Xi'an Jiaotong University between October 2016 and February 2018 was retrieved and retrospectively analyzed by WFO. The recommendations were divided into 'recommended', 'considered' and 'not recommended' groups. Results were considered concordant when oncologists' recommendations were categorized as 'recommended' or 'for consideration' by WFO. RESULTS: The concordance rate of 200 subjects with postoperative adjuvant therapy was 77%. However, the rate was 27.5% in the remaining 102 cases with metastatic disease receiving either first-line or no treatment. Further analysis demonstrated that inconsistencies were mainly due to different choices of chemotherapy regimens. Subgroup study indicates that tumor stage, receptor status and age also had influences at the concordance rate. CONCLUSION: The results of this study suggest that WFO is a promising artificial intelligence system for the treatment of breast cancer. These findings can also serve as a reference framework for the inclusion of artificial intelligence in the ongoing medical reform in China.

Transcranial Magnetic Stimulation for Pain, Headache, and Comorbid Depression: INS-NANS Expert Consensus Panel Review and Recommendation.

BACKGROUND: While transcranial magnetic stimulation (TMS) has been studied for the treatment of psychiatric disorders, emerging evidence supports its use for pain and headache by stimulating either motor cortex (M1) or dorsolateral prefrontal cortex (DLPFC). However, its clinical implementation is hindered due to a lack of consensus in the quality of clinical evidence and treatment recommendation/guideline(s). Thus, working collaboratively, this multinational multidisciplinary expert panel aims to: 1) assess and rate the existing outcome evidence of TMS in various pain/headache conditions; 2) provide TMS treatment recommendation/guidelines for the evaluated conditions and comorbid depression; and 3) assess the cost-effectiveness and technical issues relevant to the long-term clinical implementation of TMS for pain and headache. METHODS: Seven task groups were formed under the guidance of a 5-member steering committee with four task groups assessing the utilization of TMS in the treatment of Neuropathic Pain (NP), Acute Pain, Primary Headache Disorders, and Posttraumatic Brain Injury related Headaches (PTBI-HA), and remaining three assessing the treatment for both pain and comorbid depression, and the cost-effectiveness and technological issues relevant to the treatment. RESULTS: The panel rated the overall level of evidence and recommendability for clinical implementation of TMS as: 1) high and extremely/strongly for both NP and PTBI-HA respectively; 2) moderate for postoperative pain and migraine prevention, and recommendable for migraine prevention. While the use of TMS for treating both pain and depression in one setting is clinically and financially sound, more studies are required to fully assess the long-term benefit of the treatment for the two highly comorbid conditions, especially with neuronavigation. CONCLUSIONS: After extensive literature review, the panel provided recommendations and treatment guidelines for TMS in managing neuropathic pain and headaches. In addition, the panel also recommended more outcome and cost-effectiveness studies to assess the feasibility of the long-term clinical implementation of the treatment.

Subgroup identification via homogeneity pursuit for dense longitudinal/spatial data.

In the clinical trial community, it is usually not easy to find a treatment that benefits all patients since the reaction to treatment may differ substantially across different patient subgroups. The heterogeneity of treatment effect plays an essential role in personalized medicine. To facilitate the development of tailored therapies and improve the treatment efficacy, it is important to identify subgroups that exhibit different treatment effects. We consider a very general framework for subgroup identification via the homogeneity pursuit methods usually employed in econometric time series analysis. The change point detection algorithm in our procedure is most suitable for analyzing dense longitudinal or spatial data which are quite common for biomedical studies these days. We demonstrate that our proposed method is fast and accurate through extensive numerical studies. In particular, our method is illustrated by analyzing a diffusion tensor imaging data set.

Estimating individualized optimal combination therapies through outcome weighted deep learning algorithms.

With the advancement in drug development, multiple treatments are available for a single disease. Patients can often benefit from taking multiple treatments simultaneously. For example, patients in Clinical Practice Research Datalink with chronic diseases such as type 2 diabetes can receive multiple treatments simultaneously. Therefore, it is important to estimate what combination therapy from which patients can benefit the most. However, to recommend the best treatment combination is not a single label but a multilabel classification problem. In this paper, we propose a novel outcome weighted deep learning algorithm to estimate individualized optimal combination therapy. The Fisher consistency of the proposed loss function under certain conditions is also provided. In addition, we extend our method to a family of loss functions, which allows adaptive changes based on treatment interactions. We demonstrate the performance of our methods through simulations and real data analysis.

Predicting Therapy Success and Costs for Personalized Treatment Recommendations Using Baseline Characteristics: Data-Driven Analysis.

BACKGROUND: Different treatment alternatives exist for psychological disorders. Both clinical and cost effectiveness of treatment are crucial aspects for policy makers, therapists, and patients and thus play major roles for healthcare decision-making. At the start of an intervention, it is often not clear which specific individuals benefit most from a particular intervention alternative or how costs will be distributed on an individual patient level. OBJECTIVE: This study aimed at predicting the individual outcome and costs for patients before the start of an internet-based intervention. Based on these predictions, individualized treatment recommendations can be provided. Thus, we expand the discussion of personalized treatment recommendation. METHODS: Outcomes and costs were predicted based on baseline data of 350 patients from a two-arm randomized controlled trial that compared treatment as usual and blended therapy for depressive disorders. For this purpose, we evaluated various machine learning techniques, compared the predictive accuracy of these techniques, and revealed features that contributed most to the prediction performance. We then combined these predictions and utilized an incremental cost-effectiveness ratio in order to derive individual treatment recommendations before the start of treatment. RESULTS: Predicting clinical outcomes and costs is a challenging task that comes with high uncertainty when only utilizing baseline information. However, we were able to generate predictions that were more accurate than a predefined reference measure in the shape of mean outcome and cost values. Questionnaires that include anxiety or depression items and questions regarding the mobility of individuals and their energy levels contributed to the prediction performance. We then described how patients can be individually allocated to the most appropriate treatment type. For an incremental cost-effectiveness threshold of 25,000 €/quality-adjusted life year, we demonstrated that our recommendations would have led to slightly worse outcomes (1.98%), but with decreased cost (5.42%). CONCLUSIONS: Our results indicate that it was feasible to provide personalized treatment recommendations at baseline and thus allocate patients to the most beneficial treatment type. This could potentially lead to improved decision-making, better outcomes for individuals, and reduced health care costs.

Treatment recommendations within the leeway of clinical guidelines: A qualitative interview study on oncologists' clinical deliberation.

BACKGROUND: Recommending the optimal treatment for an individual patient requires a well-balanced consideration of various medical, social and ethical factors. The interplay of these factors, interpretation of the patient's situation and understanding of the existing clinical guidelines can lead to divergent therapy recommendations, depending on the attending physician. Gaining a better understanding of the individual process of medical decision-making and the differences occurring will support the delivery of optimal individualized care within the clinical setting. METHODS: A case vignette of a 64-year-old patient with locally advanced pancreatic adenocarcinoma was discussed with oncologists in 14 qualitative, semi-structured interviews at two academic institutions. Relevant factors that emerged were ranked by the participants using the Q card sorting method. Qualitative data analysis and descriptive statistics were performed. RESULTS: Oncologists recommend different therapeutic approaches within the leeway of the relevant clinical guidelines. One group of participants endorses a rather aggressive and potentially curative approach with a combination chemotherapy following the FOLFIRINOX protocol to provide the patient with the best chances of resectability. The second group suggests a milder chemotherapy approach with gemcitabine, highlighting the palliative approach and the patient's quality of life. Clinical guidelines are generally seen as an important point of reference, but are complicated to apply in highly individual cases. CONCLUSION: The physician's individual assessment of factors, such as biological age, general condition or prognosis, plays a decisive role in treatment recommendations, particularly in those cases which are not fully covered by guidelines. Judgment and discretion remain crucial in clinical decision-making and cannot and should not be fully ruled out by evidence-based guidelines. Therefore, a more comprehensive reflection on the interaction between evidence-based medicine and the physician's estimation of each individual case is desirable. Knowledge of existing barriers can enhance the implementation of guidelines, for example, through medical education.

Adjuvant treatment recommendations for patients with ER-positive/HER2-negative early breast cancer by Swiss tumor boards using the 21-gene recurrence score (SAKK 26/10).

BACKGROUND: To evaluate the effect of Recurrence Score® results (RS; Oncotype DX® multigene assay ODX) on treatment recommendations by Swiss multidisciplinary tumor boards (TB). METHODS: SAKK 26/10 is a multicenter, prospective cohort study of early breast cancer patients: Eligibility: R0-resection, ≥10% ER+ malignant cells, HER2-, pN0/pN1a. Patients were stratified into low-risk (LR) and non-low-risk (NLR) groups based on involved nodes (0 vs 1-3) and five additional predefined risk factors. Recommendations were classified as hormonal therapy (HT) or chemotherapy plus HT (CT + HT). Investigators were blinded to the statistical analysis plan. A 5%/10% rate of recommendation change in LR/NLR groups, respectively, was assumed independently of RS (null hypotheses). RESULTS: Two hundred twenty two evaluable patients from 18 centers had TB recommendations before and after consideration of the RS result. A recommendation change occurred in 45 patients (23/154 (15%, 95% CI 10-22%) in the LR group and 22/68 (32%, 95% CI 22-45%) in the NLR group). In both groups the null hypothesis could be rejected (both p < 0.001). Specifically, in the LR group, only 5/113 (4%, 95% CI 1-10%) with HT had a recommendation change to CT + HT after consideration of the RS, while 18/41 (44%, 95% CI 28-60%) of patients initially recommended CT + HT were subsequently recommended only HT. In the NLR group, 3/19 (16%, 95% CI 3-40%) patients were changed from HT to CT + HT, while 19/48 (40%, 95% CI 26-55%) were changed from CT + HT to HT. CONCLUSION: There was a significant impact of using the RS in the LR and the NLR group but only 4% of LR patients initially considered for HT had a recommendation change (RC); therefore these patients could forgo ODX testing. A RC was more likely for NLR patients considered for HT. Patients considered for HT + CT have the highest likelihood of a RC based on RS.

Treatment recommendations for schizophrenia, major depression and alcohol dependence and stigmatizing attitudes of the public: results from a German population survey.

In addition to mental health literacy, several potentially conflicting emotions and attitudes among the public are hypothesized to guide their recommendations for specific mental health treatments. It is unclear whether evidence-based treatment strategies are guided by pro-social or stigmatizing attitudes and emotions. In a representative population survey in Germany (n = 3642), we asked respondents to what extent they would recommend psychotropic medication, psychotherapy and relaxation techniques for a person with mental illness described in an unlabelled vignette. For each treatment recommendation, we used multinomial logistic regression analyses to obtain predicted probabilities. Predictors comprised illness recognition, vignette condition, causal beliefs (current stress, childhood adversities, biogenetic), emotions (fear, anger, pro-social reactions), social distance, age, gender and education. Fear predicted greater probability for recommending psychotropic drugs in all investigated illnesses (p < 0.001), whereas associations of fear with recommending psychotherapy were generally lower and no associations with the recommendation for relaxation techniques were found. Anger was related to fewer recommendations for psychotherapy in all illnesses (p < 0.01). Pro-social reactions were predominantly related to the recommendation of relaxation techniques for a person with schizophrenia or major depression (p < 0.001). Higher desire for social distance predicted fewer recommendations for relaxation techniques in all three vignette conditions (p < 0.05). Our study corroborates findings that treatment recommendations are not necessarily linked to pro-social reactions or mental health literacy. The recommendation for a treatment modality like psychotropic medication or psychotherapy can be linked to underlying fear, possibly reflecting a public desire for protection against people with mental illness.

Managing risk and patient involvement in choosing treatment for cancer: an analysis of two communication practices.

Drawing on conversation analyses of oncology consultations collected in Italy, the article examines the communication practices used to recommend treatments. We found that the oncologist formulates the treatment recommendation (TR) for high-risk patients in terms of a 'mandatory' choice and for low-risk patients as an 'optional' type of decision. In the first case the doctor presses to reach a decision during the visit while in the second case leaves the decision open-ended. Results show that high-risk patients have less time to decide, are pressured towards choosing an option, but have more opportunities for involvement in TR during the visit. Low-risk patients instead have more time and autonomy to make a choice, but they are also less involved in the decision-making in the visit time. Moreover, we document that TR is organised through sequential activities in which the oncologist informs the patient of alternative therapeutic options while at the same time building a case for the kind of treatment she/he believes to be best for the patient's health. We suggest that in this field risk plays a key role in decision-making which should be better understood with further studies and taken into account in the debate on shared decision-making and patient-centred communication.

Appropriate osteoporosis treatment by family physicians in response to FRAX vs CAROC reporting: results from a randomized controlled trial.

Canadian guidelines recommend either the FRAX or the Canadian Association of Radiologists and Osteoporosis Canada (CAROC) fracture risk assessment tools to report 10-yr fracture risk as low (<10%), moderate (10%-20%) or high (>20%). It is unknown whether one reporting system is more effective in helping family physicians (FPs) identify individuals who require treatment. Individuals ≥50 yr old with a distal radius fracture and no previous osteoporosis diagnosis or treatment were recruited. Participants underwent a dual-energy x-ray absorptiometry scan and answered questions about fracture risk factors. Participants' FPs were randomized to receive either a FRAX report or the standard CAROC-derived bone mineral density report currently used by the institution. Only the FRAX report included statements regarding treatment recommendations. Within 3 mo, all participants were asked about follow-up care by their FP, and treatment recommendations were compared with an osteoporosis specialist. Sixty participants were enrolled (31 to FRAX and 29 to CAROC). Kappa statistics of agreement in treatment recommendation were 0.64 for FRAX and 0.32 for bone mineral density. The FRAX report was preferred by FPs and resulted in better postfracture follow-up and treatment that agreed more closely with a specialist. Either the clear statement of fracture risk or the specific statement of treatment recommendations on the FRAX report may have supported FPs to make better treatment decisions.

Individualized survival prediction and surgery recommendation for patients with glioblastoma.

Background: There is a lack of individualized evidence on surgical choices for glioblastoma (GBM) patients. Aim: This study aimed to make individualized treatment recommendations for patients with GBM and to determine the importance of demographic and tumor characteristic variables in the selection of extent of resection. Methods: We proposed Balanced Decision Ensembles (BDE) to make survival predictions and individualized treatment recommendations. We developed several DL models to counterfactually predict the individual treatment effect (ITE) of patients with GBM. We divided the patients into the recommended (Rec.) and anti-recommended groups based on whether their actual treatment was consistent with the model recommendation. Results: The BDE achieved the best recommendation effects (difference in restricted mean survival time (dRMST): 5.90; 95% confidence interval (CI), 4.40-7.39; hazard ratio (HR): 0.71; 95% CI, 0.65-0.77), followed by BITES and DeepSurv. Inverse probability treatment weighting (IPTW)-adjusted HR, IPTW-adjusted OR, natural direct effect, and control direct effect demonstrated better survival outcomes of the Rec. group. Conclusion: The ITE calculation method is crucial, as it may result in better or worse recommendations. Furthermore, the significant protective effects of machine recommendations on survival time and mortality indicate the superiority of the model for application in patients with GBM. Overall, the model identifies patients with tumors located in the right and left frontal and middle temporal lobes, as well as those with larger tumor sizes, as optimal candidates for SpTR.

Proposing surgery at the prosthetic clinic: managing patient resistance.

OBJECTIVE: Investigating doctors' communicative practices for recommending surgery to amputees when the proposal counters patients' expectation. METHOD: Conversation Analysis of 77 videorecorded medical consultations at an Italian prosthesis clinic. RESULTS: Compared to the direct format doctors used to prescribe prosthesis, when suggesting surgery doctors adopted a more circuitous, indirect approach. They used a range of communication strategies, orientating to patients' likely resistance - indeed, patients were frequently observed to reject surgical options. CONCLUSIONS: Considering patients' expectations is part of a patient centred approach, hence the cautious ways in which doctors introduce the option of surgery. Moreover, doctors do not pursue recommending surgery when patients display their reluctance or resistance. PRACTICE IMPLICATIONS: Doctors in prosthetics clinics might adopt a more balanced communicative strategy that takes into account patients' perspectives, concerns and expectations, whilst but also providing patients with the necessary information to collaborate meaningfully to decision making.

Artificial intelligence-driven prognostic system for conception prediction and management in intrauterine adhesions following hysteroscopic adhesiolysis: a diagnostic study using hysteroscopic images.

Introduction: Intrauterine adhesions (IUAs) caused by endometrial injury, commonly occurring in developing countries, can lead to subfertility. This study aimed to develop and evaluate a DeepSurv architecture-based artificial intelligence (AI) system for predicting fertility outcomes after hysteroscopic adhesiolysis. Methods: This diagnostic study included 555 intrauterine adhesions (IUAs) treated with hysteroscopic adhesiolysis with 4,922 second-look hysteroscopic images from a prospective clinical database (IUADB, NCT05381376) with a minimum of 2 years of follow-up. These patients were randomly divided into training, validation, and test groups for model development, tuning, and external validation. Four transfer learning models were built using the DeepSurv architecture and a code-free AI application for pregnancy prediction was also developed. The primary outcome was the model's ability to predict pregnancy within a year after adhesiolysis. Secondary outcomes were model performance which evaluated using time-dependent area under the curves (AUCs) and C-index, and ART benefits evaluated by hazard ratio (HR) among different risk groups. Results: External validation revealed that using the DeepSurv architecture, InceptionV3+ DeepSurv, InceptionResNetV2+ DeepSurv, and ResNet50+ DeepSurv achieved AUCs of 0.94, 0.95, and 0.93, respectively, for one-year pregnancy prediction, outperforming other models and clinical score systems. A code-free AI application was developed to identify candidates for ART. Patients with lower natural conception probability indicated by the application had a higher ART benefit hazard ratio (HR) of 3.13 (95% CI: 1.22-8.02, p = 0.017). Conclusion: InceptionV3+ DeepSurv, InceptionResNetV2+ DeepSurv, and ResNet50+ DeepSurv show potential in predicting the fertility outcomes of IUAs after hysteroscopic adhesiolysis. The code-free AI application based on the DeepSurv architecture facilitates personalized therapy following hysteroscopic adhesiolysis.

Challenges and perspectives in use of artificial intelligence to support treatment recommendations in clinical oncology.

Artificial intelligence (AI) promises to be the next revolutionary step in modern society. Yet, its role in all fields of industry and science need to be determined. One very promising field is represented by AI-based decision-making tools in clinical oncology leading to more comprehensive, personalized therapy approaches. In this review, the authors provide an overview on all relevant technical applications of AI in oncology, which are required to understand the future challenges and realistic perspectives for decision-making tools. In recent years, various applications of AI in medicine have been developed focusing on the analysis of radiological and pathological images. AI applications encompass large amounts of complex data supporting clinical decision-making and reducing errors by objectively quantifying all aspects of the data collected. In clinical oncology, almost all patients receive a treatment recommendation in a multidisciplinary cancer conference at the beginning and during their treatment periods. These highly complex decisions are based on a large amount of information (of the patients and of the various treatment options), which need to be analyzed and correctly classified in a short time. In this review, the authors describe the technical and medical requirements of AI to address these scientific challenges in a multidisciplinary manner. Major challenges in the use of AI in oncology and decision-making tools are data security, data representation, and explainability of AI-based outcome predictions, in particular for decision-making processes in multidisciplinary cancer conferences. Finally, limitations and potential solutions are described and compared for current and future research attempts.

Deep Learning-Based Surgical Treatment Recommendation and Nonsurgical Prognosis Status Classification for Scaphoid Fractures by Automated X-ray Image Recognition.

Biomedical information retrieval for diagnosis, treatment and prognosis has been studied for a long time. In particular, image recognition using deep learning has been shown to be very effective for cancers and diseases. In these fields, scaphoid fracture recognition is a hot topic because the appearance of scaphoid fractures is not easy to detect. Although there have been a number of recent studies on this topic, no studies focused their attention on surgical treatment recommendations and nonsurgical prognosis status classification. Indeed, a successful treatment recommendation will assist the doctor in selecting an effective treatment, and the prognosis status classification will help a radiologist recognize the image more efficiently. For these purposes, in this paper, we propose potential solutions through a comprehensive empirical study assessing the effectiveness of recent deep learning techniques on surgical treatment recommendation and nonsurgical prognosis status classification. In the proposed system, the scaphoid is firstly segmented from an unknown X-ray image. Next, for surgical treatment recommendation, the fractures are further filtered and recognized. According to the recognition result, the surgical treatment recommendation is generated. Finally, even without sufficient fracture information, the doctor can still make an effective decision to opt for surgery or not. Moreover, for nonsurgical patients, the current prognosis status of avascular necrosis, non-union and union can be classified. The related experimental results made using a real dataset reveal that the surgical treatment recommendation reached 80% and 86% in accuracy and AUC (Area Under the Curve), respectively, while the nonsurgical prognosis status classification reached 91% and 96%, respectively. Further, the methods using transfer learning and data augmentation can bring out obvious improvements, which, on average, reached 21.9%, 28.9% and 5.6%, 7.8% for surgical treatment recommendations and nonsurgical prognosis image classification, respectively. Based on the experimental results, the recommended methods in this paper are DenseNet169 and ResNet50 for surgical treatment recommendation and nonsurgical prognosis status classification, respectively. We believe that this paper can provide an important reference for future research on surgical treatment recommendation and nonsurgical prognosis classification for scaphoid fractures.