Preoperative Prediction Model Research Articles

A Novel Multimodal Radiomics Model for Preoperative Prediction of Lymphovascular Invasion in Rectal Cancer

Objective: To explore a new predictive model of lymphatic vascular infiltration (LVI) in rectal cancer based on magnetic resonance (MR) and computed tomography (CT).Methods: A retrospective study was conducted on 94 patients with histologically confirmed rectal cancer, they were randomly divided into training cohort (n = 65) and validation cohort (n = 29). All patients underwent MR and CT examination within 2 weeks before treatment. On each slice of the tumor, we delineated the volume of interest on T2-weighted imaging, diffusion weighted imaging, and enhanced CT images, respectively. A total of 1,188 radiological features were extracted from each patient. Then, we used the student t-test or Mann–Whitney U-test, Spearman's rank correlation and least absolute shrinkage and selection operator (LASSO) algorithm to select the strongest features to establish a single and multimodal logic model for predicting LVI. Receiver operating characteristic (ROC) curves and calibration curves were plotted to determine how well they explored LVI prediction performance in the training and validation cohorts.Results: An optimal multi-mode radiology nomogram for LVI estimation was established, which had significant predictive power in training (AUC, 0.884; 95% CI, 0.803–0.964) and validation (AUC, 0.876; 95% CI, 0.721–1.000). Calibration curve and decision curve analysis showed that the multimodal radiomics model provides greater clinical benefits.Conclusion: Multimodal (MR/CT) radiomics models can serve as an effective visual prognostic tool for predicting LVI in rectal cancer. It demonstrated great potential of preoperative prediction to improve treatment decisions.

A Radiomics Model for Preoperative Prediction of Brain Invasion in Meningioma Non-Invasively Based on MRI: A Multicentre Study

Background: Prediction of brain invasion pre-operatively rather than postoperatively would contribute to the selection of surgical techniques, predicting meningioma grading and prognosis. Here, we aimed to predict the risk of brain invasion in meningioma pre-operatively using a nomogram by incorporating both radiomic features and clinical factors. Methods: A total of 1728 patients from two centres were diagnosed with meningiomas (including WHO grades I, II, and III) by histopathological examination. Radiomic features were extracted from the T1C and T2 modalities. The least absolute shrinkage and selection operator was used to select the most informative features of different modalities. Subsequently, the support vector machine algorithm was used to predict the risk of brain invasion. Furthermore, a nomogram was constructed by incorporating radiomics signature and clinical risk factors, and decision curve analysis was used to validate the clinical usefulness of the nomogram. Findings: Sixteen features were significantly correlated with brain invasion. The clinicoradiomic model derived from the fusing MRI sequences and sex resulted in the best discrimination ability for risk prediction of brain invasion, with AUCs of 0•857 (95% CI, 0•831-0•887) and 0•819 (95% CI, 0•775-0•863) and sensitivities of 72•8% (95% CI, 67•3%-78•4%) and 90•1% (95% CI, 83•4%-96•9%) in the training and validation cohorts, respectively. Interpretation: The clinicoradiomic model incorporating the fusing radiomic features and sex showed good performance and high sensitivity for risk prediction of brain invasion in meningioma, and can be applied in patients with meningiomas. Funding Statement: This work was supported by the National Natural Science Foundation of China (81772006). Declaration of Interests: The authors declare no potential conflicts of interest. Ethics Approval Statement: For this retrospective analysis, ethical approval was obtained from the Institutional Review Board of Lanzhou University Second Hospital and Beijing Tiantan Hospital, and informed consent of patients were waived for this retrospective study.

CT-based radiomics and machine learning to predict spread through air space in lung adenocarcinoma

Spread through air space (STAS) is a novel invasive pattern of lung adenocarcinoma and is also a risk factor for recurrence and worse prognosis of lung adenocarcinoma. The aims of this study are to develop and validate a computed tomography (CT)‑based radiomics model for preoperative prediction of STAS in lung adenocarcinoma. This retrospective study was approved by an institutional review board and included 462 (mean age, 58.06years) patients with pathologically confirmed lung adenocarcinoma. STAS was identified in 90 patients (19.5%). Two experienced radiologists segmented and extracted radiomics features on preoperative thin-slice CT images with radiomics extension independently. Intraclass correlation coefficients (ICC) and Pearson's correlation were used to rule out those low reliable (ICC < 0.75) and redundant (r> 0.9) features. Univariate logistic regression was applied to select radiomics features which were associated with STAS. A radiomics-based machine learning predictive model using a random forest (RF) was developed and calibrated with fivefold cross-validation. The diagnostic performance of the model was measured by the area under the curve (AUC) of receiver operating characteristic (ROC). With univariate analysis, 12 radiomics features and age were found to be associated with STAS significantly. The RF model achieved an AUC of 0.754 (a sensitivity of 0.880 and a specificity of 0.588) for predicting STAS. CT-based radiomics model can preoperatively predict STAS in lung adenocarcinoma with good diagnosis performance. • CT-based radiomics and machine learning model can predict spread through air space (STAS) in lung adenocarcinoma with high accuracy. • The random forest (RF) model achieved an AUC of 0.754 (a sensitivity of 0.880 and a specificity of 0.588) for predicting STAS.

Preoperative prediction of postoperative cerebellar mutism syndrome. Validation of existing MRI models and proposal of the new Rotterdam pCMS prediction model

PurposePostoperative cerebellar mutism syndrome (pCMS) is a complication that may occur after pediatric fossa posterior tumor surgery. Liu et al. developed an MRI-based prediction model to estimate pCMS risk preoperatively. The goal of this study was to validate the model of Liu et al. and if validation was not as sensitive in our group as previously described to develop an easy to use, reliable, and sensitive preoperative risk prediction model for pCMS.MethodsIn this study, 121children with a fossa posterior tumor who underwent surgery at ErasmusMC/Sophia Children’s Hospital, the Netherlands between 2004 and 2018 could be included. Twenty-six percent of them developed pCMS. Preoperative MRI were scored using the Liu et al. model.ResultsThe Liu et al. model reached an accuracy of 78%, a sensitivity of 58%, and a specificity of 84% in our cohort. In a new risk model some of the variables of Liu et al. were included as well as some of the recently described preoperative MRI characteristics in pCMS patients by Zhang et al. The new model reached an accuracy of 87%, a sensitivity of 97%, and a specificity of 84% in our patient group.ConclusionBecause the Liu et al. model did not provide an as accurate risk prediction in our cohort as was expected, we created a new risk prediction model that reached high model accuracy in our cohort that could assist neurosurgeons in determining their surgical tactics and help prepare high risk patients and their parents for this severe complication.

Abstract P3-07-04: A dedicated nomogram to predict nodal pathological complete response in node-positive breast cancer patients undergoing neoadjuvant chemotherapy

Abstract Background and rationale In breast cancer patients with nodal metastases at presentation, neoadjuvant chemotherapy (NAC) may downstage axillary disease. Until a few years ago, axillary lymph node dissection (ALND) has been the standard of care for node positive disease, irrespective of nodal status after completion of NAC. However, nodal pathologic complete response (pCR) is well documented in about 40-70% of women presenting with cN+ status at baseline. In this subset of patiens, ALND is therefore unnecessary. Sentinel lymph node (SLN) biopsy is accepted as a staging procedure for clinically node negative patients at baseline. Recent trials have assessed the feasibility of SLN biopsy also for patients who had clinical conversion of nodal status from cN+ to ycN0, if at least three SLNs are retrieved. However, a higher false-negative rate is the major point of controversy about adoption of SLN biopsy in these patients. A preoperative predictive model, able to assess the likelihood of axillary pCR after NAC, could help to select those patients who might be suitable candidates for SLN surgery. This approach could provide reliable nodal staging information, avoiding unjustified ALNDs and associated morbidities. Methods A retrospective review of all node-positive breast cancer patients treated by NAC between November 2000 and April 2019 at the Breast Unit of IRCCS Maugeri Hospital was performed. Patients were considered as cN+ in case of palpable axillary lymph nodes and/or suspicious lymphadenopathy on ultrasound and/or biopsy-proven nodal metastasis. After NAC completion, all patients have been treated by ALND. Baseline pre-NAC and post-NAC variables were collected and analyzed in a multivariate analysis to find predictors of axillary pCR. Identified predictors were included to develop a dedicated nomogram. Results A total of 371 clinically node-positive patients were identified. Axillary pCR (ypN0 status) was achieved in 142 patients (38%). In multivariate analysis, post-NAC clinical N stage (ycN0 vs ycN+, OR 4.61, 95%CI 2.73-7.78, p&amp;lt;0.0001), biomolecular subtype (for triple-negative breast cancer: OR 4.36, 95%CI 1.45-13.15, p=0.009; for ER+/HER2+: OR 3.68, 95%CI 1.86-7.3, p=0.0002; for ER-/HER2+: OR 3.03, 95%CI 1.06-8.62, p=0.038)-) and clinical complete response on the breast assessed on ultrasound (OR 2.26, 95%CI 1.11-4.59, p=0.024) resulted to be the strongest indipendent predictors of axillary pCR. Based on statistical findings and clinically relevant factors, a nomogram for prediction of pCR after NAC was developed and included the above-mentioned variables and Ki67, grading and pre-NAC clinical T stage. The developed model demonstrated a good discrimination capacity (AUC: 0.77). Conclusions The developed nomogram could help to identify, among patiens cN+ at time of diagnosis, those more likely to achieve nodal pCR after completing NAC. In clinical practice, this tool could contribute to select suitable cadidates for SLN biopsy, preventing them from unnecessary ALND. Citation Format: Fabio Corsi, Valentina Forlini, Daniela Bossi, Sara Albasini, Marta Truffi, Luca Sorrentino. A dedicated nomogram to predict nodal pathological complete response in node-positive breast cancer patients undergoing neoadjuvant chemotherapy [abstract]. In: Proceedings of the 2019 San Antonio Breast Cancer Symposium; 2019 Dec 10-14; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2020;80(4 Suppl):Abstract nr P3-07-04.

Noninvasive CT radiomic model for preoperative prediction of lymph node metastasis in early cervical carcinoma

To build and validate a CT radiomic model for pre-operatively predicting lymph node metastasis in early cervical carcinoma. A data set of 150 patients with Stage IB1 to IIA2 cervical carcinoma was retrospectively collected from the Nanfang hospital and separated into a training cohort (n = 104) and test cohort (n = 46). A total of 348 radiomic features were extracted from the delay phase of CT images. Mann-Whitney U test, recursive feature elimination, and backward elimination were used to select key radiomic features. Ridge logistics regression was used to build a radiomic model for prediction of lymph node metastasis (LNM) status by combining radiomic and clinical features. The area under the receiver operating characteristic curve (AUC) and κ test were applied to verify the model. Two radiomic features from delay phase CT images and one clinical feature were associated with LNM status: log-sigma-2-0 mm-3D_glcm_Idn (p = 0.01937), wavelet-HL_firstorder_Median (p = 0.03592), and Stage IB (p = 0.03608). Radiomic model was built consisting of the three features, and the AUCs were 0.80 (95% confidence interval: 0.70 ~ 0.90) and 0.75 (95% confidence intervalI: 0.53 ~ 0.93) in training and test cohorts, respectively. The κ coefficient was 0.84, showing excellent consistency. A non-invasive radiomic model, combining two radiomic features and a International Federation of Gynecology and Obstetrics stage, was built for prediction of LNM status in early cervical carcinoma. This model could serve as a pre-operative tool. A noninvasive CT radiomic model, combining two radiomic features and the International Federation of Gynecology and Obstetrics stage, was built for prediction of LNM status in early cervical carcinoma.

National incidence, mortality outcomes, and predictors of spinal cord ischemia after thoracic endovascular aortic repair

ObjectiveSpinal cord ischemia (SCI) is a dreaded complication of thoracic endovascular aortic repair (TEVAR). There are limited national data describing the incidence and influence of in-hospital SCI recovery on survival. Moreover, no robust preoperative SCI risk assessment models currently exist. The purpose of this analysis was to analyze the Vascular Quality Initiative to determine the national incidence, survival association, and preoperative predictors of SCI after TEVAR. MethodsAll Vascular Quality Initiative TEVAR procedures (June 2014-June 2019) were reviewed. The primary end point was development of in-hospital SCI, defined as any new neurologic deficit or paralysis not attributable to intracranial disease. Secondary end points were disease-specific SCI rates and long-term out-of-hospital survival. Functional outcomes (transient vs permanent SCI) were independently determined by treating physicians. Kaplan-Meier analysis and Cox proportional hazards methodology were used to assess the association of SCI with survival. A logistic regression model of candidate preoperative SCI predictors was created, and bootstrapped backward elimination (retaining predictors with ≥50% selection frequency) was used for model reduction. Model fit and performance statistics were validated by adjustment for Efron's optimism. ResultsThe overall rate of SCI was 3.7% (n = 422/11,473; transient, 1.6% [n = 179]; permanent, 2.1% [n = 243]). Patients who developed any SCI had significantly lower Kaplan-Meier survival estimate compared with those without SCI (1-year survival: SCI, 65%; no SCI, 87%; P < .0001), and patients with permanent SCI had notably worse survival than patients with transient SCI (1-year survival: permanent SCI, 54%; transient SCI, 80%; P < .0001). Disease-specific incidence of any SCI was as follows: aneurysm, 3.4%; dissection, 5.3%; aneurysm from dissection, 4.1%; trauma, 1.1%; penetrating ulceration, 2.4%; intramural hematoma, 5.7%; penetrating ulceration and intramural hematoma, 4.3%; and aortic thrombus, 4.8%. Several factors were selected on multivariable analysis as the most robust preoperative predictors of any SCI, including distal landing zone 5 to zone 10, nonelective case, creatinine concentration >1.38 mg/dL, smoking history, American Society of Anesthesiologists class, adjunctive procedure, nonwhite race, and preoperative hypertension (area under the curve = 0.72; Nagelkerke R2 = 0.06). ConclusionsSCI is a devastating complication after TEVAR that is associated with worse overall survival, particularly when no functional recovery occurs by hospital discharge. Disease-specific, real-world benchmarks for SCI rates are provided that may inform quality initiatives focused on reducing this complication. Importantly, this analysis is the first description of a preoperative prediction model derived from national data for determining SCI risk after TEVAR. These predictors should be used to identify high-risk patients to balance the risk of SCI and its associated increased short-term mortality with the risk of the underlying disease. Furthermore, all available adjunctive measures should be implemented in these high-risk patients to reduce risk of SCI.

MRI-based radiomics model for preoperative prediction of 5-year survival in patients with hepatocellular carcinoma

BackgroundRecurrence is the major cause of mortality in patients with resected HCC. However, without a standard approach to evaluate prognosis, it is difficult to select candidates for additional therapy.MethodsA total of 201 patients with HCC who were followed up for at least 5 years after curative hepatectomy were enrolled in this retrospective, multicentre study. A total of 3144 radiomics features were extracted from preoperative MRI. The random forest method was used for radiomics signature building, and five-fold cross-validation was applied. A radiomics model incorporating the radiomics signature and clinical risk factors was developed.ResultsPatients were divided into survivor (n = 97) and non-survivor (n = 104) groups based on the 5-year survival after surgery. The 30 most survival-related radiomics features were selected for the radiomics signature. Preoperative AFP and AST were integrated into the model as independent clinical risk factors. The model demonstrated good calibration and satisfactory discrimination, with a mean AUC of 0.9804 and 0.7578 in the training and validation sets, respectively.ConclusionsThis radiomics model is a valid method to predict 5-year survival in patients with HCC and may be used to identify patients for clinical trials of perioperative therapies and for additional surveillance.

A Preoperative Risk Prediction Model for Lymph Node Examination of Stage I-III Colon Cancer Patients: A Population-Based Study.

Background: Lymph node examination is a prognostic indicator for colon cancer (CC) patients. The aim of this study was to develop and validate a preoperative risk prediction model for inadequate lymph node examination.Methods: 24284 patients diagnosed as stage I-III CC between 2010-2014 were extracted from SEER database and randomly divided into development cohort (N=12142) and internal validation cohort (N=12142). 680 patients diagnosed as stage I-III CC between 2012-2014 were extracted from our hospital as external validation cohort. Logistic regression analysis was performed and risk score of each factor was calculated according to model formula. Model discrimination was assessed using C-statistics.Results: Preoperative risk factors were identified as gender, age, tumor site and tumor size. Patients with total risk score of 0-6 were considered as low risk group while patients scored ≥13 were considered as high risk group. The model had good discrimination and calibration in all cohorts and could apply to patients in the SEER database (American population) and patients in our hospital (Chinese population).Conclusions: The model could accurately predict the risk of inadequate lymph node examination before surgery and might provide useful reference for surgeons and pathologists.

Protocol for the derivation and external validation of a 30-day mortality risk prediction model for older patients having emergency general surgery (PAUSE score—Probability of mortality Associated with Urgent/emergent general Surgery in oldEr patients score)

IntroductionPeople 65 years and older represent the fastest growing segment of the surgical population. Older age is associated with doubling of risk when undergoing emergency general surgery (EGS) procedures and...

A Diagnostic Nomogram of Pathologic Grade for Preoperative Risk Stratification in Upper Tract Urothelial Carcinoma.

Background:To develop a novel nomogram to improve the preoperative diagnosis of pathological grade of upper tract urothelial carcinoma (UTUC).Methods:Retrospective study was conducted with 245 patients with UTUC treated by radical nephroureterectomy from 2002 to 2016. Of the cohort, 57.6% received ureteroscopic (URS) biopsy and 35.9% received urine cytology examination. Preoperative clinical characteristics and examination results were collected. Final pathological grade was diagnosed by postoperative pathology. Univariable and multivariable binary logistic regressions were applied to establish a preoperative predictive model for tumor grade, and significant factors were included in the nomogram. The area under curve (AUC) was used to show the predictive efficacy, and the calibration plot was drawn for validation.Results:Of the 245 patients, 72.7% were diagnosed with pathological high-grade disease. Age (odds ratio [OR] = 1.03, P = .039), sessile (OR = 3.86, P = .021), positive urinary cytology (OR = 6.87, P = .035), and biopsy high-grade result (OR = 10.85, P < .001) were independent predictors for pathological high-grade disease. The predictive nomogram containing these factors achieved an AUC of 0.78, which was significantly better than URS biopsy alone (AUC = 0.62, P = .003) in the whole cohort. In the URS biopsy subgroup, the nomogram achieved an AUC of 0.79, better than biopsy alone (AUC = 0.76), but was not statistically significant (P = .431). When the cutoff value of the nomogram was set at 0.64, the sensitivity of detecting a high-grade lesion versus low-grade lesion was 80.3%, better than that of URS biopsy alone (sensitivity = 65.7%).Conclusions:Advanced age, sessile, positive urinary cytology, and biopsy high-grade were independent predictors of pathological high-grade disease in patients with UTUC. A nomogram containing these factors can improve diagnostic accuracy, potentially reducing the risk of “undergrading” by URS biopsy.

The Role of Preoperative Dynamic Contrast-enhanced 3.0-T MR Imaging in Predicting Early Recurrence in Patients With Early-Stage Hepatocellular Carcinomas After Curative Resection.

Objectives: Liver resection is potentially curative for early-stage hepatocellular carcinoma (eHCC) in patients with well-preserved liver function. The prognosis of these patients after resection is still unsatisfactory because of frequent early recurrence (ER). Therefore, we investigated the role of preoperative dynamic contrast-enhanced 3.0-T MR imaging in predicting ER of eHCC after curative resection.Methods From May 2014 to October 2017, we retrospectively analyzed 82 patients with eHCC who underwent dynamic MR imaging and subsequently underwent curative resection. Liver Imaging Reporting and Data System (LI-RADS) v2018 major and ancillary imaging features, as well as two non-LI-RADS MR imaging features (irregular tumor margin and tumor number), were evaluated. A multivariate Cox regression analysis was used to identify independent predictors, and two models (preoperative and postoperative prediction models) were developed.Results ER was observed in 25 patients (25/82, 30.5%). In the univariate analyses, preoperative alpha-fetoprotein (AFP) level >200 ng/ml, three MR imaging features (multifocal tumors, corona enhancement, and irregular tumor margin), and microvascular invasion (MVI) were associated with ER. In the multivariate analysis, corona enhancement (hazard ratio [HR]: 2.970; p = 0.013) and irregular tumor margin (HR: 2.377; p = 0.048) were independent predictors in the preoperative prediction model, and preoperative AFP level >200 ng/ml (HR: 2.493; p = 0.044) plus corona enhancement (HR: 3.046; p = 0.014) were independent predictors in the postoperative prediction model (microvascular invasion [MVI] was not; p = 0.061). When combined with both predictors, the specificity for ER in the preoperative prediction model was 98.2% (56/57), which was comparable to that of the postoperative prediction model [96.7% (55/57)].Conclusions Our results demonstrated that preoperative MR imaging features (corona enhancement and irregular tumor margin) have the potential to preoperatively identify high-risk ER patients with eHCC, with a specificity >90%.

Preoperative prediction model of anastomotic leakage after colorectal surgery

Objective To analyze the preoperative risk factors of anastomotic leakage after colorectal surgery, construct a preoperative prediction model of anastomotic leakage after colorectal surgery, evaluate the probability of anastomotic leakage after surgery, and guide clinical decision-making. Methods From January 1, 2007 to December 31, 2016, 1 340 patients who underwent colorectal surgery and primary anastomosis in the Central Theater Command General Hospital of the Chinese People’s Liberation Army (formerly Wuhan General Hospital of Guangzhou Military Region) were collected. The risk factors of anastomotic leakage were analyzed by R-3.5.2. The forecasting model of contour map was drawn by RMS software package in R software. Results There are 123 cases of anastomotic leakage occurred in this study. The incidence of anastomotic leakage was 9.18% (123/1 340). Preoperative body mass index (BMI) was larger (Z=2.128, P 0.75 indicated that the model had good predictive ability. The AUC for predicting anastomotic leakage after colorectal surgery is 0.86, which has good predictive performance. Conclusion Large BMI, high NRS2002 score, no nutritional support at the lesion site in rectum and before operation are independent risk factors for anastomotic leakage. The constructed contour map prediction model is helpful to predict the probability of anastomotic leakage after operation. Key words: Colorectal; Anastomotic leakage; Risk factors; Prediction model

A Novel Preoperative Nomogram for Predicting Lymph Node Invasion in Renal Cell Carcinoma Patients Without Metastasis.

ObjectiveTo provide a preoperative predictive model to support clinical decision-making regarding the selection of in renal cell carcinoma (RCC) patients who will benefit the most from lymph node dissection.MethodsThis retrospective analysis enrolled 374 RCC patients without distant metastasis who underwent surgical treatment from January 2006 to December 2017. The relationships between lymph node invasion (LNI) and age at surgery; gender; body mass index(BMI); the presence of clinical symptoms such as flank pain, hematuria or a palpable mass; clinical T stage (cT stage); clinical N stage (cN stage); and the results of routine hematological and serum biochemical analyses were investigated. All the variables were included in univariate and multivariate logistic regression analyses, and the significant variables were then included in a novel nomogram to predict the probability of LNI. Then, we calibrated the nomogram with an internal validation set.ResultsSix of eighteen variables were significant in the univariate logistic regression analysis. After multivariate logistic regression analysis, age at surgery (OR=0.643, 95% CI: 0.421–0.975), cT stage (OR=3.034, 95% CI: 1.541–5.926), cN stage (OR=6.353, 95% CI: 3.273–12.456), lymphocyte percentage (OR=0.481, 95% CI: 0.256–0.894), and the presence of clinical symptoms (OR=2.045, 95% CI: 1.065–3.924) were independent predictors of LNI and were included in the nomogram. The C-index of this nomogram was 0.824.ConclusionPreoperative basic laboratory findings combined with the results of a physical examination and radiological examination can indicate the probability of LNI in RCC patients.

Development and assessment of machine learning algorithms for predicting remission after transsphenoidal surgery among patients with acromegaly.

Preoperative prediction of transsphenoidal surgical (TSS) response is important for determining individual treatment strategies for acromegaly. There is currently no accurate predictive model for TSS response for acromegaly. The current study sought to develop and validate machine learning (ML)-based models for preoperative prediction of TSS response for acromegaly. Six hundred sixty-eight patients with acromegaly were enrolled and divided into training (n = 534) and text datasets (n = 134) in this retrospective, data mining and ML study. The forward search algorithm was used to select features, and six ML algorithms were applied to construct TSS response prediction models. The performance of these ML models was validated using receiver operating characteristics analysis. Model calibration, discrimination ability, and clinical usefulness were also assessed. Three hundred forty-nine (52.2%) patients achieved postoperative remission criteria and exhibited good TSS response. A univariate analysis was conducted and eight features, including age, hypertension, ophthalmic disorders, GH, IGF-1, nadir GH, maximal tumor diameter, and Knosp grade, were significantly associated with the TSS response in patients with acromegaly. After feature selection, the gradient boosting decision tree (GBDT), which was constructed with the eight significant features showed the best favorable discriminatory ability both the training (AUC = 0.8555) and validation (AUC = 0.8178) cohorts. The GBDT model showed good discrimination ability and calibration, with the highest levels of accuracy and specificity, and provided better estimates of TTS responses of patients with acromegaly compared with using only the Knosp grade. Decision curve analysis confirmed that the model was clinically useful. ML-based models could aid neurosurgeons in the preoperative prediction of TTS response for patients with acromegaly, and could contribute to determining individual treatment strategies.

Preoperative radiographic clues for transdural disc herniation: could it be predictable?

Transdural disc herniation (TDH) is a rare event accounting for 0.3-1.5% of all disc herniation cases. Considering the risk of leakage of the cerebrospinal fluid from the dural defect after removal of TDH or incomplete removal, it is very important to recognize TDH before surgery. This study is a retrospective case analysis to analyze the imaging findings of seven cases and to construct a preoperative prediction model for TDH. Retrospective radiographic examination was performed among patients operated for TDH in two institutions from 2008 to 2018. The radiographic images were analyzed according to the following eight signs: including absence of dural tent, complete block of spinal canal, hawk-beak sign, double-layered lesion, increased distance between the dura and cauda equina, rim enhancement, dural tent enhancement, and epidural gas. To clarify the predictive ability of these radiographic signs, consecutive 131 surgically confirmed epidural disc herniation (EDH) patients for the last 2years were set as a control group for TDH. The sum of radiographic findings was compared between TDH and EDH patients to determine the cutoff value. There were 1 thoracic and 6 lumbar TDHs among 75 thoracic and 6674 lumbar disc herniation cases with an incidence of 1.33% and 0.09%, respectively. Dural tent (p = 0.000, odds ratio = 106.67), double-layered lesion (p = 0.000, odds ratio = 22.69), and distance between the dura and cauda equina (p = 0.007, odds ratio = 52.00) were statistically significantly different between TDH and EDH. According to the receiver operating characteristic curve, the cutoff value of 1.5 had 85.7% sensitivity and 90.8% specificity. Preoperative imaging can be useful for TDH diagnosis. It is safe to consider the possibility of TDH in patients with more than two findings in the preoperative images.

Development of Preoperative Prediction Models for Pain and Functional Outcome After Total Knee Arthroplasty Using The Dutch Arthroplasty Register Data

BackgroundOne of the main determinants of treatment satisfaction after total knee arthroplasty (TKA) is the fulfillment of preoperative expectations. For optimal expectation management, it is useful to accurately predict the treatment result. Multiple patient factors registered in the Dutch Arthroplasty Register (LROI) can potentially be utilized to estimate the most likely treatment result. The aim of the present study is to create and validate models that predict residual symptoms for patients undergoing primary TKA for knee osteoarthritis. MethodsData were extracted from the LROI of all TKA patients who had preoperative and postoperative patient-reported outcome measures registered. Multivariable logistic regression analyses were performed to construct predictive algorithms for satisfaction, treatment success, and residual symptoms concerning pain at rest and during activity, sit-to-stand movement, stair negotiation, walking, performance of activities of daily living, kneeling, and squatting. We assessed predictive performance by examining measures of calibration and discrimination. ResultsData of 7071 patients could be included for data analysis. Residual complaints on kneeling (female 72%/male 59%) and squatting (female 71%/male 56%) were reported most frequently, and least residual complaints were scored for walking (female 16%/male 12%) and pain at rest (female 18%/male 14%). The predictive algorithms were presented as clinical calculators that present the probability of residual symptoms for an individual patient. The models for residual symptoms concerning sit-to-stand movement, stair negotiation, walking, activities of daily living, and treatment success showed acceptable discriminative values (area under the curve 0.68-0.74). The algorithms for residual complaints regarding kneeling, squatting, pain, and satisfaction showed less favorable results (area under the curve 0.58-0.64). The calibration curves showed adequate calibration for most of the models. ConclusionA considerable proportion of patients have residual complaints after TKA. The present study showed that demographic and patient-reported outcome measure data collected in the LROI can be used to predict the probability of residual symptoms after TKA. The models developed in the present study predict the chance of residual symptoms for an individual patient on 10 specific items concerning treatment success, functional outcome, and pain relief. This prediction can be useful for individualized expectation management in patients planned for TKA.

Computed Tomography Radiomic Nomogram for Preoperative Prediction of Extrathyroidal Extension in Papillary Thyroid Carcinoma

Objectives: Determining the presence of extrathyroidal extension (ETE) is important for patients with papillary thyroid carcinoma (PTC) in selecting the proper surgical approaches. This study aimed to explore a radiomic model for preoperative prediction of ETE in patients with PTC.Methods: The study included 624 PTC patients (without ETE, n = 448; with minimal ETE, n = 52; with gross ETE, n = 124) whom were divided randomly into training (n = 437) and validation (n = 187) cohorts; all data were gathered between January 2016 and November 2017. Radiomic features were extracted from computed tomography (CT) images of PTCs. Key radiomic features were identified and incorporated into a radiomic signature. Combining the radiomic signature with clinical risk factors, a radiomic nomogram was constructed using multivariable logistic regression. Delong test was used to compare different receiver operating characteristic curves.Results: Five key radiomic features were incorporated into the radiomic signature, which were significantly associated with ETE (p < 0.001 for both cohorts) and slightly better than clinical model integrating significant clinical risk factors in the training cohort (area under the receiver operating characteristic curve (AUC), 0.791 vs. 0.778; F1 score, 0.729 vs. 0.714) and validation cohort (AUC, 0.772 vs. 0.756; F1 score, 0.710 vs. 0.692). The radiomic nomogram significantly improved predictive value in the training cohort (AUC, 0.837, p < 0.001; F1 score, 0.766) and validation cohort (AUC, 0.812, p = 0.024; F1 score, 0.732).Conclusions: The radiomic nomogram significantly improved the preoperative prediction of ETE in PTC patients. It indicated that radiomics could be a valuable method in PTC research.

Preoperative predictive model and nomogram for disease recurrence following radical nephroureterectomy for high grade upper tract urothelial carcinoma

PurposeTo identify preoperative risk factors for disease recurrence, following radical nephroureterectomy (RNU) for upper tract urothelial carcinoma (UTUC), and to create a predictive nomogram. Materials and methodsBased on a multicenter database, we identified patients who underwent RNU due to high grade UTUC. Urothelial carcinoma of the bladder or contralateral UTUC was not considered as recurrence. Cox regression model was used to determine the effect of different preoperative variables as predictors of recurrence. ResultsTwo hundred and forty-five patients were included in the analysis. The 2 and 5 years recurrence rates were 16.3% and 19.2%, respectively. Factors associated with recurrence on univariable analysis were sessile architecture hazard ratio (HR) 3.16 (95% CI, 1.38–7.26, P = 0.006), ≥cT3 disease HR 2.30 (95% CI, 1.12–4.72, P= 0.023), age >65 HR 2.02 (95% CI, 1.00–4.05, P= 0.048), Eastern Cooperative Group > 0 HR 1.98 (95% CI, 1.09–3.57, P= 0.023), hydronephrosis HR 1.93 (95% CI, 1.04–3.57, P= 0.035). Higher hemoglobin levels HR 0.81 (95% CI, 0.69–0.96, P= 0.013) and preoperative estimated glomerular filtration rate ≥ 50 HR 0.48 (95% CI, 0.25–0.92, P = 0.028) were associated with lower probability for recurrence. Multivariable analysis identified sessile architecture as the only independent predictor of recurrence HR 2.52 (95% CI, 1.09–5.86, P= 0.0308). C-index of 0.71 was calculated for a predictive model including all variables in the multivariable analysis, indicating good predictive accuracy. A nomogram predicting 2 and 5 year recurrence free probability was developed accordingly. ConclusionsBased on a multicenter database, we developed a nomogram with good predictive accuracy for recurrence following RNU. This may serve as an aid in decision-making regarding the use of neoadjuvant chemotherapy.

Development of predictive models for all individual questions of SRS-22R after adult spinal deformity surgery: a step toward individualized medicine

Health-related quality of life (HRQL) instruments are essential in value-driven health care, but patients often have more specific, personal priorities when seeking surgical care. The Scoliosis Research Society-22R (SRS-22R), an HRQL instrument for spinal deformity, provides summary scores spanning several health domains, but these may be difficult for patients to utilize in planning their specific care goals. Our objective was to create preoperative predictive models for responses to individual SRS-22R questions at 1 and 2years after adult spinal deformity (ASD) surgery to facilitate precision surgical care. Two prospective observational cohorts were queried for ASD patients with SRS-22R data at baseline and 1 and 2years after surgery. In total, 150 covariates were used in training machine learning models, including demographics, surgical data and perioperative complications. Validation was accomplished via an 80%/20% data split for training and testing, respectively. Goodness of fit was measured using area under receiver operating characteristic (AUROC) curves. In total, 561 patients met inclusion criteria. The AUROC ranged from 56.5 to 86.9%, reflecting successful fits for most questions. SRS-22R questions regarding pain, disability and social and labor function were the most accurately predicted. Models were less sensitive to questions regarding general satisfaction, depression/anxiety and appearance. To the best of our knowledge, this is the first study to explicitly model the prediction of individual answers to the SRS-22R questionnaire at 1 and 2years after deformity surgery. The ability to predict individual question responses may prove useful in preoperative counseling in the age of individualized medicine. These slides can be retrieved under Electronic Supplementary Material.