Low-rank fusion convolutional neural network for prediction of remission after stereotactic radiosurgery in patients with acromegaly: a proof-of-concept study.

Artificial intelligence approaches to analyze pathological images (pathomic) for outcome prediction have not been sufficiently considered in the field of pituitary research. A total of 5,504 hematoxylin & eosin-stained pathology image tiles from 58 acromegalic patients with a good or poor outcome were integrated with other clinical and genetic information to train a low-rank fusion convolutional neural network (LFCNN). The model was externally validated in 1,536 patches from an external cohort. The primary outcome was the time to the first endocrine remission after stereotactic radiosurgery (SRS). The median time of initial endocrine remission was 43 months (interquartile range [IQR]: 13-60 months) after SRS, and the 24-month initial cumulative remission rate was 57.9% (IQR: 46.4-72.3%). The patient-wise accuracy of the LFCNN model in predicting the primary outcome was 92.9% in the internal test dataset, and the sensitivity and specificity were 87.5 and 100.0%, respectively. The LFCNN model was a strong predictor of initial cumulative remission in the training cohort (hazard ratio [HR] 9.58, 95% confidence interval [CI] 3.89-23.59; p < 0.001) and was higher than that of established prognostic markers. The predictive value of the LFCNN model was further validated in an external cohort (HR 9.06, 95% CI 1.14-72.25; p = 0.012). In this proof-of-concept study, clinically and genetically useful prognostic markers were integrated with digital images to predict endocrine outcomes after SRS in patients with active acromegaly. The model considerably outperformed established prognostic markers and can potentially be used by clinicians to improve decision-making regarding adjuvant treatment choices. © 2022 The Pathological Society of Great Britain and Ireland.

Radiomics analysis allows for precise prediction of silent corticotroph adenoma among non-functioning pituitary adenomas.

OBJECTIVE: To predict silent corticotroph adenomas (SCAs) among non-functioning pituitary adenomas preoperatively using noninvasive radiomics. METHODS: A total of 302 patients including 146 patients diagnosed with SCAs and 156 patients with non-SCAs were enrolled (training set: n = 242; test set: n = 60). Tumor segmentation was manually generated using ITK-SNAP. From T1-weighted imaging (T1WI), T2-weighted imaging (T2WI), and contrast-enhanced T1WI, 2550 radiomics features were extracted using Pyradiomics. Pearson's correlation coefficient values were calculated to exclude redundant features. Several machine learning algorithms were developed to predict SCAs incorporating the radiomics and semantic features including clinical, laboratory, and radiology-associated features. The performance of models was evaluated by AUC. RESULTS: Patients in the SCA group were younger (49.5 vs 55.2 years old) and more female (85.6% vs 37.2%) than those in the non-SCA group (p < 0.001). More invasiveness (p = 0.011) and cystic and microcystic change (p < 0.001) were observed in patients with SCAs. The ensemble algorithm presented the largest AUC of 0.927 among all the algorithms trained in the test set, and the accuracy, specificity, and sensitivity of predicting SCAs were all 0.867 (at cut-off 0.5). The overall model performed better than that only using semantic features available in the clinic. Radiomics prediction was the most important feature, with gender ranking second and age ranking third. Radiomics features on T2WI were superior to those on other MR modalities in SCA prediction. CONCLUSION: Our ensemble learning model outperformed current clinical practice in differentiating patients with SCAs and non-SCAs using radiomics, which might help make appropriate treatment strategies. KEY POINTS: • Radiomics might improve the preoperative diagnosis of SCAs by MR images. • T2WI was superior to T1WI and CE-T1WI in the preoperative diagnosis of SCAs. • The ensemble machine learning model outperformed current clinical practice in SCAs diagnosis and treatment decision-making could be more individualised using the nomogram.

Machine learning in predicting early remission in patients after surgical treatment of acromegaly: a multicenter study.

PURPOSE: Accurate prediction of postoperative remission is beneficial for effective patient-physician communication in acromegalic patients. This study aims to train and validate machine learning prediction models for early endocrine remission of acromegalic patients. METHODS: The training cohort included 833 patients with growth hormone (GH) secreting pituitary adenoma from 2010 to 2018. We trained a partial model (only using pre-operative variables) and a full model (using all variables) to predict off-medication endocrine remission at six-month follow-up after surgery using multiple algorithms. The models were validated in 99 prospectively collected patients from a second campus and 52 patients from a third institution. RESULTS: C-statistic and the accuracy of the best partial model was 0.803 (95% CI 0.757-0.849) and 72.5% (95% CI 67.6-77.5%), respectively. C-statistic and the accuracy of the best full model was 0.888 (95% CI 0.861-0.914) and 80.3% (95% CI 77.5-83.1%), respectively. The c-statistics (and accuracy) of using only Knosp grade, total resection, or postoperative day 1 GH level as the single predictor were lower than our partial model or full model (p < 0.001). C-statistics remained similar in the prospective cohort (partial model 0.798, and full model 0.903) and in the external cohort (partial model 0.771, and full model 0.871). A web-based application integrated with the trained models was published at  https://deepvep.shinyapps.io/Acropred/ . CONCLUSION: We developed and validated interpretable and applicable machine learning models to predict early endocrine remission after surgical resection of a GH-secreting pituitary adenoma. Predication accuracy of the trained models were better than those using single variables.

COMPARATIVE EFFICACY OF MEDICAL TREATMENT FOR ACROMEGALY: A SYSTEMATIC REVIEW AND NETWORK META-ANALYSIS OF INTEGRATED RANDOMIZED TRIALS AND OBSERVATIONAL STUDIES.

Objective: Comprehensive evidence comparing different medications for acromegaly is scarce. The aim of this study was to perform a network meta-analysis based on evidence from both randomized trials and observational studies of medical treatments for acromegaly. Methods: Electronic databases were searched for both observational studies and randomized trials that enrolled acromegaly patients treated with medications of interest. Simulated trials were generated by a machine learning algorithm and then synthesized with Bayesian random-effects network meta-analyses. The main outcome was the rate of insulin-like growth factor 1 (IGF-1) control after medical treatment. Results: We included 90 studies (100 arms, 4,523 patients) before matching. After matching, 28 simulated trials were generated. Balance of matched arms was checked by spatial distance and correlation matrix. Cotreatment with somatostatin receptor ligands and pegvisomant was the most effective treatment compared with other treatments. In unselected patients, pegvisomant was better than octreotide long-acting release (logOR, 0.85; 95% credible interval [CrI], 0.05 to 1.65) or lanreotide (logOR, 1.09, 95% CrI, 0.05 to 2.14), and the mean absolute IGF-1 control rate ranged from 40 to 60%. In partially responsive patients, cotreatment with somatostatin receptor ligands and pegvisomant was similar to pegvisomant monotherapy, ranking as the most two effective treatments, and the mean absolute IGF-1 control rate was over 60%. Conclusion: Our analysis suggested that the combination of data from observational studies and randomized trials in network meta-analysis was feasible. The findings of this network meta-analysis provided robust evidence supporting the current guidelines in treatment strategy for acromegaly. Abbreviations: CrI = credible interval; DA = dopamine agonist; GH = growth hormone; IGF-1 = insulin-like growth factor 1; ITT = intention-to-treat; LAN = lanreotide; LAN-ATG = lanreotide autogel; OCT = octreotide; OCT-LAR = octreotide long acting repeatable; OR = odds ratio; PEG = pegvisomant; PP = per-protocol; SRL = somatostatin receptor ligand.

Surgical outcomes and multidisciplinary management strategy of Cushing's disease: a single-center experience in China.

OBJECTIVE: The primary aim of this study was to investigate the value of multidisciplinary team (MDT) management in treating patients with Cushing's disease (CD). The secondary aim was to assess the concordance of bilateral inferior petrosal sinus sampling (BIPSS) lateralization with intraoperative observations. METHODS: The authors recruited 124 consecutive patients (128 procedures) who had undergone endoscopic endonasal resection of adrenocorticotropic hormone-secreting pituitary adenomas from May 2014 to April 2018 and assessed their clinical characteristics, surgical outcomes, and adjuvant therapies. The criteria for surgical remission were normalized serum and urinary cortisol levels, which could be suppressed by a low-dose dexamethasone suppression test at 3-months' follow-up without adjuvant treatment. RESULTS: The remission rates of the 113 patients with long-term follow-up (20.3 ± 12.2 months) were 83.2% after surgery alone and 91.2% after adjuvant therapy. The surgical remission rates of macroadenomas, MRI-visible microadenomas, and MRI-negative tumors were 66.7% (12/18), 89.3% (67/75), and 75% (15/20), respectively (p = 0.039). The surgical remission rates had a trend of improvement during the study period (87.5% in 2017-2018 vs 76.5% in 2014, p = 0.517). Multivariate regression analysis showed that a history of previous pituitary surgery (OR 0.300, 95% CI 0.100-0.903; p = 0.032) and MRI-visible microadenoma (OR 3.048, 95% CI 1.030-9.019; p = 0.044) were independent factors influencing surgical remission. The recurrence rate was 3.2% after a mean of 18 months after surgery. The remission rate of postoperative MDT management in patients with persistent disease was higher than non-MDT management (66.7% vs 0%, p = 0.033). In cases with preoperative BIPSS lateralization, 84.6% (44/52) were concordant with intraoperative findings. CONCLUSIONS: MRI-visible microadenoma and primary surgery were independent predictors of surgical remission in CD. The MDT management strategy helps to achieve a better overall outcome. BIPSS may help to lateralize the tumor in MRI-negative/equivocal microadenomas.

Cushing's disease in older patients: Presentation and outcome.

BACKGROUND: To define the symptoms, signs and treatment outcomes in a population of older patients with Cushing's disease (CD). METHODS: We analysed the clinical presentation and treatment outcomes in 45 CD patients older than 60 years, in comparison with 90 CD patients younger than 60, and a control group of 45 older patients with nonfunctioning pituitary adenomas. We reviewed preoperative clinical characteristics, medical comorbidities, imaging findings and endocrine testing as well as surgical and endocrine outcomes. RESULTS: Older CD patients had significantly lower body mass index (BMI) (P = 0.031), were more likely to have muscle wasting (P = 0.006) and women were less likely to have hirsutism (P = 0.033). Older patients with CD had more medical comorbidities than younger patients, which correlated with a higher ASA grade (P < 0.001), but the surgical complication rates were similar in both groups. Surgical remission was achieved in 38/45 (84.4%) older patients and 78/90 (86.7%) younger patients (P = NS). Recurrent disease was more frequent in younger patients (19.2%) in comparison with older patients (2.6%, P = 0.019). CONCLUSIONS: Older patients with CD appear to have a distinct phenotype with a more catabolic picture, including a lower BMI and greater prevalence of muscle wasting. Surgical outcomes are similar without a significant age-related increase in complications.

Predictive value of T2 relative signal intensity for response to somatostatin analogs in newly diagnosed acromegaly.

INTRODUCTION: The difficulty of predicting the efficacy of somatostatin analogs (SSA) is not fully resolved. Here, we quantitatively evaluated the predictive value of relative signal intensity (rSI) on T1- and T2-weighted magnetic resonance imaging (MRI) for the short-term efficacy (3 months) of SSA therapy in patients with active acromegaly and assessed the correlation between MRI rSI and expression of somatostatin receptors (SSTR). METHODS: This was a retrospective review of prospectively recorded data. Ninety-two newly diagnosed patients (37 males and 55 females) with active acromegaly were recruited. All patients were treated with pre-surgical SSA, followed by reassessment and transspenoidal surgery. rSI values were generated by calculating the ratio of SI in the tumor to the SI of normal frontal white matter. The Youden indices were calculated to determine the optimal cutoff of rSI to determine the efficacy of SSA. The correlation between rSI and expression of SSTR2/5 was analyzed by the Spearman rank correlation coefficient. RESULTS: T2 rSI was strongly correlated with biochemical sensitivity to SSA. The cutoff value of T2 rSI to distinguish biochemical sensitivity was 1.205, with a positive predictive value (PPV) of 81.5 % and a negative predictive value (NPV) of 77.3 %. No correlation was found between MRI and tumor size sensitivity. Moreover, T2 rSI was negatively correlated with the expression of SSTR5. CONCLUSION: T2 rSI correlates with the expression of SSTR5 and quantitatively predicts the biochemical efficacy of SSA in acromegaly.

Retinal nerve fiber layer changes after transsphenoidal and transcranial pituitary adenoma resection.

PURPOSE: Retinal nerve fiber layer (RNFL) will show retrograde degeneration following damage to the optic nerve or the optic tract in patients with pituitary adenoma. RNFL changes after surgery have not been studied thoroughly in patients with the transsphenoidal surgery and patients with the transcranial surgery. METHODS: Thirty-seven patients with pituitary adenoma were recruited from Huashan hospital between September 2010 and July 2014. Patients were divided into two groups: the transsphenoidal group and the transcranial group. Before surgery, 3 and 9 months after surgery, follow-up optic coherence tomography were conducted. RESULTS: Twenty-one patients underwent transsphenoidal surgery and 16 patients underwent transcranial surgery. No obvious difference were observed between these two groups before surgery. The mean RNFL thickness did not change significantly in patients who underwent transsphenoidal surgery: 91.1 before surgery, 92.7 at 3 months after surgery (p = 0.392) and 92.8 at 9 months after surgery( p = 0.395). The mean RNFL thickness decreased inpatients who underwent transcranial surgery: 93.6 before surgery, 86.1 at 3 months after surgery (p = 0.000) and 88.1 at 9 months after surgery (p = 0.005). CONCLUSIONS: In the short time follow-up, there was no change of RNFL thickness in pituitary adenoma patients underwent transsphenoidal surgery, but a decrease inpatients underwent transcranial surgery.

Comparison of multifocal visual evoked potential, static automated perimetry, and optical coherence tomography findings for assessing visual pathways in patients with pituitary adenomas.

BACKGROUND: There have been no studies investigating the correlation between structural [thickness of the retinal nerve fiber layer (RNFL) as determined by optical coherence tomography (OCT)] and functional [Humphrey visual field (HVF) or visual evoked potential (VEP) amplitude] measures of optic nerve integrity in patients with pituitary adenomas (PA). METHODS: Patients with PAs were recruited between September 2010 and September 2013. OCT, standard automated perimetry (SAP), and multifical VEP (mfVEP) were performed. Agreement between OCT, SAP, and mfVEP values in classifying eyes/quadrants was determined using AC1 statistics. Pearson's correlation was used to examine relationships between structural and functional data. RESULTS: In total, 88.7% of the eyes tested showed abnormal SAP findings and 93.7% showed abnormal mfVEP findings. Only 14.8% of the eyes showed abnormal OCT findings. The agreement between SAP and mfVEP findings was 88.9% (AC1 = 0.87). The agreement between OCT and mfVEP findings was 24.2% (AC1 = -0.52), and that between OCT and SAP findings was 21.5% (AC1 = -0.56). The correlation values between RNFL thickness and the functional measurements were -0.601 for the mfVEP score (P = 0.000) and -0.441 for the SAP score (P = 0.000). The correlation between the mfVEP and SAP scores was -0.617 (P = 0.000). CONCLUSIONS: mfVEP, SAP, and OCT provided complementary information for detecting visual pathway abnormalities in patients with PAs. Good agreement was demonstrated between SAP and mfVEP and quantitative analysis of structure-function measurements revealed a moderate correlation.

A systematic review and meta-analysis of surgeries performed for treating deep-seated cerebral cavernous malformations.

OBJECTIVE: The clinical benefit of surgery for treatment of deep-seated cerebral cavernous malformations (CCMs) is still a matter of debate. Although the surgical removal of CCMs is widely accepted, the benefits of reducing the rate of haemorrhage must be balanced against the risk of peri-operative morbidity. Here, we provide a systematic review and meta-analysis of the clinical benefits of surgery for treating deeply localised CCMs. METHODS: A comprehensive search of PubMed and Embase was conducted to identify relevant studies. The rate and a 95% confidence interval (CI) were used to measure the risk of haemorrhage and adverse outcomes. RESULTS: A total of 34 cohort studies reporting surgeries on CCMs were included in our analysis. Overall, the average post-surgical haemorrhage rate was 1.0% (95% CI: 0.7-1.4%). Nine per cent (95% CI: 6.9-11.3%) of the patients developed adverse events at follow-up following the surgical resection of deep-seated CCMs. The percentage of transient neural defects following surgical resection was 34.6% (95% CI: 29.4-39.9%). The proportions of transient focal neurological defect before and after the year 2006 were 44.9% (95% CI: 34.1-55.8%) and 30.3% (95% CI: 25.1-35.9%), respectively. CONCLUSIONS: Our meta-analysis demonstrates post-surgical haemorrhage rate and complications related to surgeries on deep-seated CCMs. The post-surgical haemorrhage rate was low with a relatively high rate of post-surgical complications.

Presurgical Medical Treatment in Prolactinomas: Surgical Implications and Pathological Characteristics From 290 Cases.

OBJECTIVE: To review experience regarding the treatment of prolactinomas by endoscopic endonasal surgery focusing on the association between presurgical dopamine agonist (DA) treatment and perioperative outcomes, surgical morbidities, endocrine outcomes, and pathological characteristics. METHODS: A single-center series of 290 cases was analyzed retrospectively and clinical data were collected. Intratumoral collagen content was assessed by Masson trichrome staining. RESULTS: Tenacious tumor consistency (27.8% vs 9.8%, P < .001) was more common in DA-pretreated patients compared with patients who underwent initial surgery. Moreover, DA-pretreated macroadenomas presented more intraoperative blood loss (200 [100-400] mL vs 175 [100-300] mL; P = .014), longer surgical duration (177 ± 95 minutes vs 154 ± 57 minutes; P = .043), and more surgical morbidities (19.4% vs 8.9%; P = .034). Additionally, DA-pretreated macroadenomas presented a higher collagen volume fraction than that of the initial surgery group (23.6 ± 2.2% vs 13.2 ± 2.1%; P = .001). Correlation analysis revealed a close correlation between collagen volume fraction and the cumulative dose of bromocriptine (BRC) in macroadenomas (r = 0.438, P < .001). Regarding endocrine outcomes, DA-pretreated microadenomas showed a lower proportion of initial remission compared with patients who underwent initial surgery (86.7% vs 100%, P = .047). CONCLUSION: This study described increased surgical difficulty and inferior endocrine outcomes associated with tumor fibrosis secondary to presurgical BRC treatment in prolactinomas. Neurosurgeons should note that presurgical BRC treatment may render subsequent surgery more challenging.

Combined simultaneous transsphenoidal and transcranial regimen improves surgical outcomes in complex giant pituitary adenomas: A longitudinal retrospective cohort study.

BACKGROUND: Surgical treatment of complex giant pituitary adenomas (GPAs) presents significant challenges. The efficacy and safety of combining transsphenoidal and transcranial approaches for these tumors remain controversial. In this largest cohort of patients with complex GPAs, we compared the surgical outcomes between those undergoing a combined regimen and a non-combined regimen. We also examined the differences in risks of complications, costs, and logistics between the two groups, which might offer valuable information for the appropriate management of these patients. MATERIALS AND METHODS: This was a multicenter retrospective cohort study conducted at 13 neurosurgical centers. Consecutive patients who received a combined or non-combined regimen for complex GPAs were enrolled. The primary outcome was gross total resection, while secondary outcomes included complications, surgical duration, and relapse. A propensity score-based weighting method was used to account for differences between the groups. RESULTS: Out of 647 patients (298 [46.1%] women, mean age: 48.5 ± 14.0 years) with complex GPAs, 91 were in the combined group and 556 were in the non-combined group. Compared with the non-combined regimen, the combined regimen was associated with a higher probability of gross total resection (50.5% vs. 40.6%, odds ratio [OR]: 2.18, 95% confidence interval [CI]: 1.30-3.63, P = 0.003). The proportion of patients with life-threatening complications was lower in the combined group than in the non-combined group (4.4% vs. 11.2%, OR: 0.25, 95% CI: 0.08-0.78, P = 0.017). No marked differences were found between the groups in terms of other surgical or endocrine-related complications. However, the combined regimen exhibited a longer average surgery duration of 1.3 h (P < 0.001) and higher surgical costs of 22,000 CNY (approximate 3,000 USD, P = 0.022) compared with the non-combined approach. CONCLUSIONS: The combined regimen offered increased rates of total resection and decreased incidence of life-threatening complications, which might be recommended as the first-line choice for these patients.

Severe hypernatremia during postoperative care in patients with craniopharyngioma.

Purpose: We aimed to describe and predict the risk of severe hypernatremia after surgical resection of craniopharyngioma and to identify the association of water intake, urine output, and sodium level change in the patients. Method: The outcome was postoperative severe hypernatremia. We identified risk factors associated with hypernatremia using multivariable regression. We trained machine learning models to predict the outcome. We compared serum sodium change, intravenous input, oral input, total input, urine output, and net fluid balance according to different nurse shifts. Results: Among 234 included patients, 125 developed severe hypernatremia after surgery. The peak incidence occurred during day 0 and day 6 after surgery. The risk was increased in patients with gross total resection (odds ratio (OR) 2.41, P < 0.001), high Puget classification (OR 4.44, P = 0.026), preoperative adrenal insufficiency (OR 2.01, P = 0.040), and preoperative hypernatremia (OR 5.55, P < 0.001). The random forest algorithm had the highest area under the receiver operating characteristic curve (0.770, 95% CI, 0.727-0.813) in predicting the outcome and was validated in the prospective validation cohort. Overnight shifts were associated with the highest serum sodium increase (P = 0.010), less intravenous input (P < 0.001), and less desmopressin use (P < 0.001). Conclusion: The overall incidence of severe hypernatremia after surgical resection of craniopharyngioma was significant, especially in patients with gross total resection, hypothalamus distortion, preoperative adrenal insufficiency, and preoperative severe hypernatremia. Less intravenous input and less desmopressin use were associated with serum sodium increases, especially during overnight shifts.

Ectopic recurrence craniopharyngioma: series report and literature review.

BACKGROUND: Craniopharyngioma is a common intracranial tumor located in the sellar-suprasellar region. Due to the involvement of adjacent structures, it can lead to increased intracranial pressure, visual impairment, and endocrine deficiencies. Surgical resection is the primary treatment, but it is a tough challenge to achieve total resection, which will led to the frequency of recurrences and progressions. Among them, distant spread is extremely rare, but important complication, identifying and providing proper therapy, is crucial. METHODS: We report two cases of ectopic recurrence craniopharyngioma and make a literature review for the published similar case reports. RESULTS: Our literature review revealed 63 cases (including our patient). The onset age in children group and adult group ranges from 2-14 years old (6.70 ± 3.33) to 17-73 years old (40.63 ± 15.58), while the interval year between tumor initiation and ectopic recurrence ranges from 0.17-20 (7.28 ± 6.76) years to 0.3-34 (6.85 ± 7.29). Achieving gross total resection seems not to prevent the ectopic recurrence. The major pathology of ectopic recurrence craniopharyngioma is adamantinomatous type. The most common site of ectopic recurrence is frontal lobe. According to the pathogenesis, 35 cases were seeding along the surgical approach, and 28 cases were seeding via the CSF pathway. CONCLUSION: Ectopic recurrence craniopharyngioma is rare, but it can lead to serious symptoms. Delicate surgical procedure can help to reduce the risk of ectopic recurrence, and standardized follow-up can provide valuable information for treatment.

Assessment of quality of life in patients with craniopharyngioma and identification of risk factors for compromised overall wellness.

Objective: Quality of Life (QoL) has been a multifactorial concerning issue in oncology. We aimed to inspect the pre-operative QoL among patients with craniopharyngioma and to explore the potential correlations between parameters of QoL and clinical indices. Subjects and methods: We enrolled a total of 109 patients with craniopharyngioma. We utilized Short Form 36 (SF-36), Symptom Check List-90, Generalized Anxiety Disorder Questionnaire scale (GAD7), Patient Health Questionnaire Depression (PHQ9) and Pittsburgh Sleep Quality Index to prospectively evaluated their QoL. Parameters of QoL along with clinical indices were compared among sub-groups divided according to Puget classification. Correlation analyses and regression analyses were performed to detect influential determinants to self-reported wellness. Results: Patients presented impaired QoL compared with general population (p < 0.001), as assessed by SF-36. Correlation analyses indicated the detrimental influence resulting from central diabetes insipidus (CDI). Multivariate linear regression unveiled the adverse effect of CDI on Mental Component Summary (coefficient = -13.869, p= 0.007), GAD7 total score (coefficient = 2.072, p = 0.049) as well as PHQ9 total score (coefficient = 3.721, p = 0.001). Multivariate logistic regression verified CDI as a risk factor of developing depressive symptoms (OR = 6.160, p = 0.001). Conclusion: QoL of patients with craniopharyngioma was remarkably compromised before operation. CDI exerted detrimental influences on patients' QoL and it might serve as a marker for early identification of patients at risk of depression.

Machine learning prediction of venous thromboembolism after surgeries of major sellar region tumors.

PURPOSE: To describe and predict the risk of venous thromboembolism (VTE) after surgical resection of major sellar region tumors. METHOD: Patients with sellar region tumors were identified from a database. The outcome was VTE, including deep vein thrombosis (DVT) and pulmonary embolism (PE) within 60 days after surgery. We trained regression and machine learning models to predict the outcome using baseline characteristics, surgical findings and postoperative laboratory tests. RESULTS: Among 3818 patients included, 124 patients developed VTE after surgery. The total 60-day VTE incidence was 3.2 %, with incidence peak within ten days after the surgery. The risk increased in patients >65 years old (OR 2.96, p < 0.001), in patients with chordoma (OR 3.40, p = 0.006) or craniopharyngioma (OR 1.86, p = 0.036), in patients underwent craniotomy approach (OR 2.78, p = 0.017), in patients with high volume CSF leakage (OR 4.24, p < 0.001), and in patients with longer surgical duration (OR 1.78, p = 0.029). The linear discriminant analysis algorithm had the highest AUC (0.869, 95%CI, 0.840-0.898) in predicting the outcome. The specificity, accuracy, and sensitivity of the best model were 61.8 %, 93.6 %, and 92.8 %, respectively. Risk stratification using our best model suggested that 1.3 % and 24.5 % of the patient developed VTE in the low-risk group and in the high-risk group, respectively. We developed an online decision-support tool available on https://deepvep.shinyapps.io/VTEpred/. CONCLUSION: The overall incidence of VTE after surgical resection of major sellar region tumors was clinically significant, especially in older patients with chordoma or craniopharyngioma.

Development and Validation of a Clinical Prediction Model for Venous Thromboembolism Following Neurosurgery: A 6-Year, Multicenter, Retrospective and Prospective Diagnostic Cohort Study.

BACKGROUND: Based on the literature and data on its clinical trials, the incidence of venous thromboembolism (VTE) in patients undergoing neurosurgery has been 3.0%~26%. We used advanced machine learning techniques and statistical methods to provide a clinical prediction model for VTE after neurosurgery. METHODS: All patients (n = 5867) who underwent neurosurgery from the development and retrospective internal validation cohorts were obtained from May 2017 to April 2022 at the Department of Neurosurgery at the Sanbo Brain Hospital. The clinical and biomarker variables were divided into pre-, intra-, and postoperative. A univariate logistic regression (LR) was applied to explore the 67 candidate predictors with VTE. We used a multivariable logistic regression (MLR) to select all significant MLR variables of MLR to build the clinical risk prediction model. We used a random forest to calculate the importance of significant variables of MLR. In addition, we conducted prospective internal (n = 490) and external validation (n = 2301) for the model. RESULTS: Eight variables were selected for inclusion in the final clinical prediction model: D-dimer before surgery, activated partial thromboplastin time before neurosurgery, age, craniopharyngioma, duration of operation, disturbance of consciousness on the second day after surgery and high dose of mannitol, and highest D-dimer within 72 h after surgery. The area under the curve (AUC) values for the development, retrospective internal validation, and prospective internal validation cohorts were 0.78, 0.77, and 0.79, respectively. The external validation set had the highest AUC value of 0.85. CONCLUSIONS: This validated clinical prediction model, including eight clinical factors and biomarkers, predicted the risk of VTE following neurosurgery. Looking forward to further research exploring the standardization of clinical decision-making for primary VTE prevention based on this model.

Digital image analysis allows objective stratification of patients with silent PIT1-lineage pituitary neuroendocrine tumors.

Studies describing the clinical presentation and prognosis of patients with silent PIT1 (pituitary specific transcription factor)-lineage pituitary neuroendocrine tumors (PitNETs) are rare. We identified patients with positive PIT1 tumor staining but without evidence of hormone hypersecretion at a tertiary center. Clusters were obtained according to cell morphology and immunostaining from each patient's digitally segmented whole slide image. We compared the clinical presentations, radiological features, and prognoses of the different clusters. We identified 146 patients (68 male, 42.9 ± 14.1 years old) with silent PIT1-lineage PitNETs. Morphology clustering suggested that tumors with large nuclei and apparent eccentricity were associated with a higher proportion of aggressiveness and a higher hazard of recurrence [hazard ratio (HR): 2.64, (95% CI, 1.06-6.55), p = 0.037]. Immunohistochemical clustering suggested that tumors with thyroid stimulating hormone (TSH) staining or all negative PIT1-lineage hormones were associated with a higher proportion of aggressiveness and a higher risk of recurrence [HR: 12.4, (95% CI, 1.60-93.5), p = 0.015]. We obtained three-tier risk profiles by combining morphological and immunohistochemical clustering. Patients with the high-risk profile presented the highest recurrence rate compared with those in the medium-risk and low-risk profiles [HR: 3.54, (95% CI, 1.40-8.93), p = 0.002]. In conclusion, digital image analysis based on cell morphology and immunohistochemical staining allows objective stratification of patients with silent PIT1-lineage tumors. Typical morphological characteristics of high-risk tumors are large tumor nuclei and high eccentricity, and typical immunostaining characteristics are TSH staining or negative staining for all PIT1-lineage hormones.