Never tell me the odds: Investigating pro-hoc explanations in medical decision making.

This paper examines a kind of explainable AI, centered around what we term pro-hoc explanations, that is a form of support that consists of offering alternative explanations (one for each possible outcome) instead of a specific post-hoc explanation following specific advice. Specifically, our support mechanism utilizes explanations by examples, featuring analogous cases for each category in a binary setting. Pro-hoc explanations are an instance of what we called frictional AI, a general class of decision support aimed at achieving a useful compromise between the increase of decision effectiveness and the mitigation of cognitive risks, such as over-reliance, automation bias and deskilling. To illustrate an instance of frictional AI, we conducted an empirical user study to investigate its impact on the task of radiological detection of vertebral fractures in x-rays. Our study engaged 16 orthopedists in a 'human-first, second-opinion' interaction protocol. In this protocol, clinicians first made initial assessments of the x-rays without AI assistance and then provided their final diagnosis after considering the pro-hoc explanations. Our findings indicate that physicians, particularly those with less experience, perceived pro-hoc XAI support as significantly beneficial, even though it did not notably enhance their diagnostic accuracy. However, their increased confidence in final diagnoses suggests a positive overall impact. Given the promisingly high effect size observed, our results advocate for further research into pro-hoc explanations specifically, and into the broader concept of frictional AI.

Medium-long-term follow-up of slipped capital femoral epiphysis treated with in situ fixation and patient-acceptable symptom state evaluation.

This study evaluates the Patient Acceptable Symptom State (PASS) in patients with slipped capital femoral epiphysis (SCFE) treated with in situ fixation, focusing on medium to long-term outcomes and quality of life. Its primary goal is to establish a subjective well-being cutoff, using subjective methods and the iHOT33 scale, for assessing patients in future studies. Additionally, it explores functionality differences between mild and moderate-severe SCFE, case series epidemiology and potential complications. A retrospective analysis of 63 patients (73 hips), treated for SCFE between 2000 and 2017 at our facility using in situ fixation, was conducted. These patients underwent clinical, anamnestic, and radiological assessments, with PASS determined based on iHOT33 questionnaire results and statistical analysis. The mean age at surgery was 12.95 years (±1.64, range 9-17), with an average follow-up of 11 years (±4.60, range 5-20). At follow-up, 87% of patients reported achieving PASS, with higher iHOT33 scores correlating to PASS. A cutoff of >68 on the iHOT33 scale showed strong predictive ability for assessing PASS (area under the curve 0.857, 88.89% sensitivity, 79.69% specificity). The findings indicate that 87% of patients achieved PASS at medium to long-term follow-up, with better clinical function than those who did not report PASS. The iHOT33 scale's effectiveness in predicting PASS, especially with a cutoff of >68, suggests this method's efficacy. Given these positive outcomes, including in moderate-severe cases treated with in situ fixation, this approach is considered a viable therapeutic option.

Spinal Epidural Atypical Meningioma: Case Report and Review of the Literature.

Spinal atypical meningiomas are rare, and those whose main extension is in the epidural space are anecdotal. Here, we report a case of a young woman presenting with sensory disturbances and a radiological diagnosis of a dorsal epidural sleeve-like mass. The surgical resection of the lesion allowed the decompression of the spinal cord and led to the histopathological diagnosis of atypical meningioma. At the 3-month follow-up, her neurological recovery was complete. Because of the gross total removal of the lesion, adjuvant radiotherapy was not performed: At the 2-year follow-up, no recurrence of disease was detected. A comprehensive literature review was performed, and just two more case reports on epidural atypical meningiomas were found in the English literature. Through this case report and literature review, we described a rare manifestation of spinal meningioma that entered into a differential diagnosis for extradural spinal lesions, such as secondary malignancies.

Ultra-Early Treatment of Neurosurgical Emergencies with Endoscopic Endonasal Approach: Experience from Three Italian Referral Centers.

PURPOSE: the aim of this multicenter study is to preliminarily assess the role of the Endoscopic Endonasal Approach (EEA) in ultra-early (i.e., within 12 h) management of selected neurosurgical emergencies in terms of clinical and radiological outcomes. METHODS: 26 patients affected by sellar/parasellar pathologies with rapid progression of symptoms were managed with EEA within 12 h from diagnosis in three Italian tertiary referral Centers from January 2016 to December 2019. Both clinical and radiological data have been collected preoperatively as well as post-operatively in order to perform retrospective analysis. RESULTS: The average time from admission to the operating room was 5.5 h (±2.3). The extent of resection was gross-total in 20 (76.9%), subtotal in 6 (23.1%) patients. One patient experienced re-bleeding after a subtotal removal of a hemorrhagic lesion. Patients with a longer time from admission (>4 h) to the operatory room (OR) experienced stable impairment of the visual acuity (p = 0.033) and visual field (p = 0.029) in the post-operative setting. CONCLUSIONS: The Endoscopic Endonasal Approach represents a safe, effective technique that can be efficiently used with good results in the management of selected neurosurgical emergencies in centers with adequate experience.

A supervised machine-learning algorithm predicts intraoperative CSF leak in endoscopic transsphenoidal surgery for pituitary adenomas.

BACKGROUND: Despite advances in endoscopic transnasal transsphenoidal surgery (E-TNS) for pituitary adenomas (PAs), cerebrospinal fluid (CSF) leakage remains a life-threatening complication predisposing to major morbidity and mortality. In the current study we developed a supervised ML model able to predict the risk of intraoperative CSF leakage by comparing different machine learning (ML) methods and explaining the functioning and the rationale of the best performing algorithm. METHODS: A retrospective cohort of 238 patients treated via E-TNS for PAs was selected. A customized pipeline of several ML models was programmed and trained; the best five models were tested on a hold-out test and the best classifier was then prospectively validated on a cohort of 35 recently treated patients. RESULTS: Intraoperative CSF leak occurred in 54 (22,6%) of 238 patients. The most important risk's predictors were: non secreting status, older age, x-, y- and z-axes diameters, ostedural invasiveness, volume, ICD and R-ratio. The random forest (RF) classifier outperformed other models, with an AUC of 0.84, high sensitivity (86%) and specificity (88%). Positive predictive value and negative predictive value were 88% and 80% respectively. F1 score was 0.84. Prospective validation confirmed outstanding performance metrics: AUC (0.81), sensitivity (83%), specificity (79%), negative predictive value (95%) and F1 score (0.75). CONCLUSIONS: The RF classifier showed the best performance across all models selected. RF models might predict surgical outcomes in heterogeneous multimorbid and fragile populations outperforming classical statistical analyses and other ML models (SVM, ANN etc.), improving patient management and reducing preventable morbidity and additional costs.

Dissecting Stemness in Aggressive Intracranial Meningiomas: Prognostic Role of SOX2 Expression.

Meningiomas are mostly benign tumors that, at times, can behave aggressively, displaying recurrence despite gross-total resection (GTR) and progression to overt malignancy. Such cases represent a clinical challenge, particularly because they are difficult to recognize at first diagnosis. SOX2 (Sex-determining region Y-box2) is a transcription factor with a key role in stem cell maintenance and has been associated with tumorigenesis in a variety of cancers. The purpose of the present work was to dissect the role of SOX2 in predicting the aggressiveness of meningioma. We analyzed progressive/recurrent WHO grade 1−2 meningiomas and WHO grade 3 meningiomas; as controls, non-recurring WHO grade 1 and grade 2 meningioma patients were enrolled. SOX2 expression was evaluated using both immunohistochemistry (IHC) and RT-PCR. The final analysis included 87 patients. IHC was able to reliably assess SOX2 expression, as shown by the good correlation with mRNA levels (Spearman R = 0.0398, p = 0.001, AUC 0.87). SOX2 expression was an intrinsic characteristic of any single tumor and did not change following recurrence or progression. Importantly, SOX2 expression at first surgery was strongly related to meningioma clinical behavior, histological grade and risk of recurrence. Finally, survival data suggest a prognostic role of SOX2 expression in the whole series, both for overall and for recurrence-free survival (p < 0.0001 and p = 0.0001, respectively). Thus, SOX2 assessment could be of great help to clinicians in informing adjuvant treatments during follow-up.

Interpretable Machine Learning-Based Prediction of Intraoperative Cerebrospinal Fluid Leakage in Endoscopic Transsphenoidal Pituitary Surgery: A Pilot Study.

Purpose Transsphenoidal surgery (TSS) for pituitary adenomas can be complicated by the occurrence of intraoperative cerebrospinal fluid (CSF) leakage (IOL). IOL significantly affects the course of surgery predisposing to the development of postoperative CSF leakage, a major source of morbidity and mortality in the postoperative period. The authors trained and internally validated the Random Forest (RF) prediction model to preoperatively identify patients at high risk for IOL. A locally interpretable model-agnostic explanations (LIME) algorithm is employed to elucidate the main drivers behind each machine learning (ML) model prediction. Methods The data of 210 patients who underwent TSS were collected; first, risk factors for IOL were identified via conventional statistical methods (multivariable logistic regression). Then, the authors trained, optimized, and audited a RF prediction model. Results IOL reported in 45 patients (21.5%). The recursive feature selection algorithm identified the following variables as the most significant determinants of IOL: Knosp's grade, sellar Hardy's grade, suprasellar Hardy's grade, tumor diameter (on X, Y, and Z axes), intercarotid distance, and secreting status (nonfunctioning and growth hormone [GH] secreting). Leveraging the predictive values of these variables, the RF prediction model achieved an area under the curve (AUC) of 0.83 (95% confidence interval [CI]: 0.78; 0.86), significantly outperforming the multivariable logistic regression model (AUC = 0.63). Conclusion A RF model that reliably identifies patients at risk for IOL was successfully trained and internally validated. ML-based prediction models can predict events that were previously judged nearly unpredictable; their deployment in clinical practice may result in improved patient care and reduced postoperative morbidity and healthcare costs.

5-Aminolevulinic Acid (5-ALA)-Induced Protoporphyrin IX Fluorescence by Glioma Cells-A Fluorescence Microscopy Clinical Study.

5-aminolevulinic acid (5-ALA)-induced PpIX fluorescence is used by neurosurgeons to identify the tumor cells of high-grade gliomas during operation. However, the issue of whether 5-ALA-induced PpIX fluorescence consistently stains all the tumor cells is still debated. Here, we assessed the cytoplasmatic signal of 5-ALA by fluorescence microscopy in a series of human gliomas. As tumor markers, we used antibodies against collapsin response-mediated protein 5 (CRMP5), alpha thalassemia/mental retardation syndrome X-linked (ATRX), and anti-isocitrate dehydrogenase 1 (IDH1). In grade III-IV gliomas, the signal induced by 5-ALA was detected in 32.7-75.5 percent of CRMP5-expressing tumor cells. In low-grade gliomas (WHO grade II), the CRMP5-expressing tumor cells did not fluoresce following 5-ALA. Immunofluorescence with antibodies that stain various components of the blood-brain barrier (BBB) suggested that 5-ALA does not cross the un-breached BBB, in spite of its small dimension. To conclude, 5-ALA-induced PpIX fluorescence has an established role in high-grade glioma surgery, but it has limited usefulness in surgery for low-grade glioma, especially when the BBB is preserved.

Molecular Analysis in a Glioblastoma Cohort-Results of a Prospective Analysis.

The prognostic role of epidermal growth factor receptor variant III (EGFRvIII), a constitutively activated oncogenic receptor, in glioblastoma is controversial. We performed a prospective study enrolling 355 patients operated on for de novo glioblastoma at a large academic center. The molecular profile, including EGFRvIII status, MGMT promoter methylation, and VEGF expression, was assessed. Standard parameters (age, clinical status and extent of surgical resection) were confirmed to hold prognostic value. MGMT promoter methylation portended a slightly improved survival. In the whole series, confirming previous results, EGFRvIII was not associated with worsened prognosis. Interestingly, female sex was associated with a better outcome. Such findings are of interest for the design of future trials.

Machine learning model prediction of 6-month functional outcome in elderly patients with intracerebral hemorrhage.

Spontaneous intracerebral hemorrhage (ICH) has an increasing incidence and a worse outcome in elderly patients. The ability to predict the functional outcome in these patients can be helpful in supporting treatment decisions and establishing prognostic expectations. We evaluated the performance of a machine learning (ML) model to predict the 6-month functional status in elderly patients with ICH leveraging the predictive value of the clinical characteristics at hospital admission. Data were extracted by a retrospective multicentric database of patients ≥ 70 years of age consecutively admitted for the management of spontaneous ICH between January 1, 2014 and December 31, 2019. Relevant demographic, clinical, and radiological variables were selected by a feature selection algorithm (Boruta) and used to build a ML model. Outcome was determined according to the Glasgow Outcome Scale (GOS) at 6 months from ICH: dead (GOS 1), poor outcome (GOS 2-3: vegetative status/severe disability), and good outcome (GOS 4-5: moderate disability/good recovery). Ten features were selected by Boruta with the following relative importance order in the ML model: Glasgow Coma Scale, Charlson Comorbidity Index, ICH score, ICH volume, pupillary status, brainstem location, age, anticoagulant/antiplatelet agents, intraventricular hemorrhage, and cerebellar location. Random forest prediction model, evaluated on the hold-out test set, achieved an AUC of 0.96 (0.94-0.98), 0.89 (0.86-0.93), and 0.93 (0.90-0.95) for dead, poor, and good outcome classes, respectively, demonstrating high discriminative ability. A random forest classifier was successfully trained and internally validated to stratify elderly patients with spontaneous ICH into prognostic subclasses. The predictive value is enhanced by the ability of ML model to identify synergy among variables.

A Deep Learning Model for Preoperative Differentiation of Glioblastoma, Brain Metastasis and Primary Central Nervous System Lymphoma: A Pilot Study.

Background: Neuroimaging differentiation of glioblastoma, primary central nervous system lymphoma (PCNSL) and solitary brain metastasis (BM) remains challenging in specific cases showing similar appearances or atypical features. Overall, advanced MRI protocols have high diagnostic reliability, but their limited worldwide availability, coupled with the overlapping of specific neuroimaging features among tumor subgroups, represent significant drawbacks and entail disparities in the planning and management of these oncological patients. Objective: To evaluate the classification performance metrics of a deep learning algorithm trained on T1-weighted gadolinium-enhanced (T1Gd) MRI scans of glioblastomas, atypical PCNSLs and BMs. Materials and Methods: We enrolled 121 patients (glioblastoma: n=47; PCNSL: n=37; BM: n=37) who had undergone preoperative T1Gd-MRI and histopathological confirmation. Each lesion was segmented, and all ROIs were exported in a DICOM dataset. The patient cohort was then split in a training and hold-out test sets following a 70/30 ratio. A Resnet101 model, a deep neural network (DNN), was trained on the training set and validated on the hold-out test set to differentiate glioblastomas, PCNSLs and BMs on T1Gd-MRI scans. Results: The DNN achieved optimal classification performance in distinguishing PCNSLs (AUC: 0.98; 95%CI: 0.95 - 1.00) and glioblastomas (AUC: 0.90; 95%CI: 0.81 - 0.97) and moderate ability in differentiating BMs (AUC: 0.81; 95%CI: 0.70 - 0.95). This performance may allow clinicians to correctly identify patients eligible for lesion biopsy or surgical resection. Conclusion: We trained and internally validated a deep learning model able to reliably differentiate ambiguous cases of PCNSLs, glioblastoma and BMs by means of T1Gd-MRI. The proposed predictive model may provide a low-cost, easily-accessible and high-speed decision-making support for eligibility to diagnostic brain biopsy or maximal tumor resection in atypical cases.

Palliative Surgery for Drug-Resistant Epilepsy in Adult Patients. A Systematic Review of the Literature and a Pooled Analysis of Outcomes.

BACKGROUND: Several types of palliative surgery to treat drug-resistant epilepsy (DRE) have been reported, but the evidence that is available is insufficient to help physicians redirect patients with DRE to the most appropriate kind of surgery. METHODS: A systematic search in the PubMed and Scopus databases was performed according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines to compare different clinical features, outcomes, and complications of adult patients submitted to callosotomy, vagal nerve stimulation, multiple subpial transections, deep brain stimulation, or responsive neurostimulation. RESULTS: After 3447 articles were screened, 36 studies were selected, including the data of 1628 patients: 76 were treated with callosotomy, 659 were treated with vagal nerve stimulation, 416 were treated with deep brain stimulation, and 477 were treated with responsive neurostimulation. No studies including patients treated with multiple subpial transections met the inclusion criteria. The global weighted average seizure frequency reduction was 50.23%, and the global responder rate was 52.12%. There were significant differences among the palliative surgical procedures in term of clinical features of patients and epilepsy, seizure frequency reduction, and percentage of responders. Complications were differently distributed as well. CONCLUSIONS: Our analysis highlights the necessity of prospective studies, possibly randomized controlled trials, to compare different forms of palliative epilepsy surgery. Moreover, by identifying the outcome predictors associated with each technique, the best responder may be profiled for each procedure.

Comparison of combined anterior-posterior and posterior-only approaches for lumbosacral chordomas: a systematic review and meta-analysis of surgical and clinical outcomes.

Lumbosacral chordoma is a slow-growing but locally aggressive tumor, resistant to adjuvant treatments and endowed with dismal prognosis. Surgery is the mainstay of treatment but the choice of surgical approach (the posterior-only approach or the combined anterior-posterior approach) remains an open question due to the need of both pursuing a surgical radicality and preserving the neurologic function. The aim of the study was to compare the surgical and clinical outcomes of these approaches in the management of lumbosacral chordomas. A systematic review and meta-analysis in agreement with the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-analyses) guidelines of papers comparing the outcomes of the two approaches was performed. Ten papers met the inclusion criteria. The combined anterior-posterior approach was more frequently performed for tumors with an upper level beyond S2 (p = 0.012). The 5-year progression-free survival was significantly higher in posterior-only approach compared with the combined anterior-posterior approach (44.7% vs 27.1%, p = 0.049). Adjuvant radiotherapy was added more frequently after a posterior-only approach (p = 0.036) and the rate of complications was significantly lower after a posterior-only approach (p = 0.040). No significant differences in sex, age, tumor diameter, entity of resection, and overall survival were observed. Posterior-only surgical approach may be a reasonable option for lumbosacral chordoma, being associated with comparable entity of surgical resection, reduced complication rate and increased 5-year progression-free survival rate as compared with combined anterior-posterior approach.

A systematic review and meta-analysis of factors involved in bone flap resorption after decompressive craniectomy.

Decompressive craniectomy (DC) is effective in controlling increasing intracranial pressure determined by a wide range of conditions, mainly traumatic brain injury (TBI) and stroke, and the subsequent cranioplasty (CP) displays potential therapeutic benefit in terms of overall neurological function. While autologous bone flap (ABF) harvested at the time of DC is the ideal material for skull defect reconstruction, it carries several risks. Aseptic bone flap resorption (BFR) is one of the most common complications, often leading to surgical failure. The aim of our study was to systematically review the literature and carry out a meta-analysis of possible factors involved in BFR in patients undergoing ABF cranioplasty after DC. A systematic review and meta-analysis was performed in accordance with the PRISMA guidelines. Different medical databases (PubMed, Embase, and Scopus) were screened for eligible scientific reports until April 30th 2021. The following data were collected for meta-analysis to assess their role in BFR: sex, age, the interval time between DC and CP, the presence of systemic factors, the etiology determining the DC, CP surgical time, CP features, VP shunt placement, CP infection. Studies including pediatric patients or with less than 50 patients were excluded. Fifteen studies were included. There was a statistically significant increased incidence of BFR in patients with CPF > 2 compared to patients with CPF ≤ 2 (54.50% and 22.76% respectively, p = 0.010). TBI was a significantly more frequent etiology in the BFR group compared to patients without BFR (61.95% and 47.58% respectively, p < 0.001). Finally, patients with BFR were significantly younger than patients without BFR (39.12 ± 15.36 years and 47.31 ± 14.78 years, respectively, p < 0.001). The funnel plots were largely symmetrical for all the studied factors. Bone flap fragmentation, TBI etiology, and young age significantly increase the risk of bone resorption. Further studies are needed to strengthen our results and to clarify if, in those cases, a synthetic implant for primary CP should be recommended.

Slipped capital femoral epiphysis: Diagnostic pitfalls and therapeutic options.

One of the most prevalent hip pathologies that develops during adolescence is Slipped Capital Femoral Epiphysis (SCFE), and over the past few decades, its incidence has been rising. To ensure an early diagnosis and prompt intervention, orthopedic surgeons should be aware of this entity. Review of recent developments in clinical examination and imaging diagnostic procedures. The presentation includes commonly used imaging methods, slippage measurement techniques, and classification schemes that are pertinent to treatment. An overview of SCFE surgery based on pertinent study findings and knowledge gained from ongoing clinical practice. The gold standard treatment for stable SCFE cases- those in which the continuity of the metaphysis and epiphysis is preserved-is pinning in situ using a single cannulated screw without reduction. However, there are disagreements over the best course of action for stable moderate/severe SCFE. On the best surgical strategy for unstable epiphysiolysis, no universal agreement has been reached. Finding the surgical procedure that will improve the long-term outcomes of a slipped capital femoral epiphysis is the question at hand.

Artificial intelligence-enhanced intraoperative neurosurgical workflow: current knowledge and future perspectives.

INTRODUCTION: Artificial intelligence (AI) and machine learning (ML) augment decision-making processes and productivity by supporting surgeons over a range of clinical activities: from diagnosis and preoperative planning to intraoperative surgical assistance. We reviewed the literature to identify current AI platforms applied to neurosurgical perioperative and intraoperative settings and describe their role in multiple subspecialties. EVIDENCE ACQUISITION: A systematic review of the literature was conducted following the PRISMA guidelines. PubMed, EMBASE, and Scopus databases were searched from inception to December 31st, 2020. Original articles were included if they: presented AI platforms implemented in perioperative, intraoperative settings and reported ML models' performance metrics. Due to the heterogeneity in neurosurgical applications, a qualitative synthesis was deemed appropriate. The risk of bias and applicability of predicted outcomes were assessed using the PROBAST tool. EVIDENCE SYNTHESIS: Forty-one articles were included. All studies evaluated a supervised learning algorithm. A total of 10 ML models were described; the most frequent were neural networks (N.=15) and tree-based models (N.=13). Overall, the risk of bias was medium-high, but applicability was considered positive for all studies. Articles were grouped into four categories according to the subspecialty of interest: neuro-oncology, spine, functional and other. For each category, different prediction tasks were identified. CONCLUSIONS: In this review, we summarize the state-of-art applications of AI for the intraoperative augmentation of neurosurgical workflows across multiple subspecialties. ML models may boost surgical team performances by reducing human errors and providing patient-tailored surgical plans, but further and higher-quality studies need to be conducted.

Reassessing the Role of Brain Tumor Biopsy in the Era of Advanced Surgical, Molecular, and Imaging Techniques-A Single-Center Experience with Long-Term Follow-Up.

Brain biopsy is the gold standard in order to establish the diagnosis of unresectable brain tumors. Few studies have investigated the long-term outcomes of biopsy patients. The aim of this single-institution-based study was to assess the concordance between radiological and histopathological diagnoses, and the long-term patient outcome. Ninety-three patients who underwent brain biopsy in the last 5 years were analyzed. We included patients treated with stereotactically guided needle, open, and neuroendoscopic biopsies. Most patients (86%) received needle biopsy. Gliomas and primary brain lymphomas comprised 88.2% of cases. The diagnostic yield was 95.7%. Serious complication and death rates were 3.2% and 2.1%, respectively. The concordance rate between radiological and histological diagnoses was 93%. Notably, the positive predictive value of radiological diagnosis of lymphoma was 100%. Biopsy allowed specific treatment in 72% of cases. Disease-related neurological worsening was the main reason that precluded adjuvant treatment. Adjuvant treatment, in turn, was the strongest prognostic factor, since the median overall survival was 11 months with vs. 2 months without treatment (p = 0.0002). Finally, advanced molecular evaluations can be obtained on glioma biopsy specimens to provide integrated diagnoses and individually tailored treatments. We conclude that, despite the huge advances in imaging techniques, biopsy is required when an adjuvant treatment is recommended, particularly in gliomas.

The reliability and quality of online patient education videos for vestibular schwannoma.

BACKGROUND: A recent trend of looking for health-related conditions on the Internet has been described, with up 70% of searchers stating that online sources have affected their medical decision-making. Patients with vestibular schwannomas (VS) use online sources, including videos, to seek information about treatment alternatives and outcomes and surgeons experience. Our study investigates the reliability and quality of VS-related online videos. METHODS: In April 2020, a search was launched on YouTube for the key terms 'vestibular schwannoma,' 'acoustic neuroma,' 'eighth cranial nerve schwannoma,' and 'eighth cranial nerve neuroma.' Results were screened for possible inclusion. Three authors independently used the DISCERN instrument to evaluate the reliability and quality of the included videos. Factors possibly influencing popularity were investigated. RESULTS: The initial search yielded 6416 videos. 38 videos were included in the final analysis. The average DISCERN score was 2.76, indicating overall poor quality and reliability of information. Only 5% scored 4.0 or more (unbiased videos that offer evidence-supported information); 31% scored between 3.0 and 3.99, and 63% scored 2.99 or less. Videos describing symptoms or the patient's clinical presentation were slightly more popular than videos without these characteristics. Surgical videos (videos containing clips of surgical procedures) were significantly more popular than non-surgical videos (p = .024) despite being of similarly poor quality (DISCERN score 2.85 vs. 2.74, respectively). CONCLUSIONS: Available patient educational videos for VS are of mixed quality and reliability: the authors describe the strengths and pitfalls of existing YouTube videos. Considering that VS is a pathology with multiple available management modalities, and that patients' decision-making is affected by the information available on the Internet, it is of great importance that good-quality informative material be released by medical, academic, or educational institutions.

Machine Learning-Based Prediction of Early Recurrence in Glioblastoma Patients: A Glance Towards Precision Medicine.

BACKGROUND: Ability to thrive and time-to-recurrence following treatment are important parameters to assess in patients with glioblastoma multiforme (GBM), given its dismal prognosis. Though there is an ongoing debate whether it can be considered an appropriate surrogate endpoint for overall survival in clinical trials, progression-free survival (PFS) is routinely used for clinical decision-making. OBJECTIVE: To investigate whether machine learning (ML)-based models can reliably stratify newly diagnosed GBM patients into prognostic subclasses on PFS basis, identifying those at higher risk for an early recurrence (≤6 mo). METHODS: Data were extracted from a multicentric database, according to the following eligibility criteria: histopathologically verified GBM and follow-up >12 mo: 474 patients met our inclusion criteria and were included in the analysis. Relevant demographic, clinical, molecular, and radiological variables were selected by a feature selection algorithm (Boruta) and used to build a ML-based model. RESULTS: Random forest prediction model, evaluated on an 80:20 split ratio, achieved an AUC of 0.81 (95% CI: 0.77; 0.83) demonstrating high discriminative ability. Optimizing the predictive value derived from the linear and nonlinear combinations of the selected input features, our model outperformed across all performance metrics multivariable logistic regression. CONCLUSION: A robust ML-based prediction model that identifies patients at high risk for early recurrence was successfully trained and internally validated. Considerable effort remains to integrate these predictions in a patient-centered care context.