Tumor Surgery Research Articles

Intraoperative label-free tissue diagnostics using a stimulated Raman histology imaging system with artificial intelligence: An initial experience.

Accurate intraoperative tissue diagnostics could impact on decision making regarding the extent of resection (EOR) during brain tumor surgery. Stimulated Raman histology (SRH) is a label-free optical imaging method that uses different biochemical properties of tissue to generate a hematoxylin-eosin-like image and, in combination with an artificial intelligence-based image classifier, offers the opportunity to obtain rapid intraoperative tissue diagnoses. The goal of this study was to report on our initial experience with SRH to evaluate its accuracy in comparison to final tissue diagnosis. We evaluated 70 consecutive adult cases with brain tumors. We compared results of the three different SRH classifier (diagnostic, molecular and tumor/non-tumor) to the respective final histopathological result. Similarly, we evaluated the isocitrate dehydrogenase (IDH) mutations in 18 patients using SRH. Lastly, we compared SRH results of samples taken from the tumor margins with early postoperative MRI. Prediction accuracy was evaluated by logistic regression and Receiver Operator Curve (ROC) analysis. We included 19 gliomas, 9 metastases, 22 meningiomas and 14 other tumor entities. Regarding accuracy of intraoperative SRH predictions, regression analysis showed an Area Under the Curve (AUC) of 0.77 (95 % C.I. 0.64-0.89, p = 0.0008), suggesting agreement of predictions with final diagnosis. For specific tumor entities, variable accuracies were observed: The highest accuracy was obtained for meningiomas followed by high-grade glioma. IDH mutations were predicted with an AUC of 0.93 (95 % C.I. 0.88-0.98; p < 0.0001). The SRH examination of tissue samples from tumor margins corresponded with postoperative MRI in 4 out of 5 cases. Our initial experience with SRH shows that this novel imaging technique is a promising approach to obtain rapid intraoperative tissue diagnosis to guide surgical decision making based on histology and cell-density. With further refinement of AI-based automated image classification and a better integration into the surgical workflow, prediction accuracy and reliability could be improved.

3D Surgical Planning for Customized Devices in Orthopaedics: Applications in Massive Hip Reconstructions of Oncological Patients

This study investigates the morphological impact of using three-dimensional (3D) printed custom implants in surgical hip reconstruction compared to the conventional bone graft and standard size implant methods. An amount of 16 patients at the Rizzoli Orthopaedic Institute who underwent hip reconstruction surgery for tumors involving the P2 pelvis region were selected using stratified sampling. Half of them were randomly selected to receive 3D-printed implants, and the other half were selected to receive standard implants with bone grafts. Six months post-surgery, computed tomography (CT) scans were used to identify the hip joint center of rotation and to measure greater the trochanter offset and acetabular inclination angle. These CT scans were also used to construct a 3D model of the pelvis for 3D measurements. The results show no significant differences in accuracy, using Student’s T-test and Mann–Whitney U-test (p-value &gt; 0.05), between the two methods for reconstructing the hip joint center of rotation or greater trochanter offset. However, 3D-printed implants showed statistically significant greater precision, using Student’s T-test (p-value &lt; 0.05), in reconstructing the acetabular inclination angle compared to the conventional bone graft and standard-sized off-the-shelf implants. This superior precision reduces the risk of impingement of the femur implant neck with the acetabulum implant cup, which directly relates to improved implant survivorship. These findings support the continued exploration of 3D printing technology for personalized orthopedic solutions.

Influence of postoperative D-dimer evaluation and intraoperative use of intermittent pneumatic vein compression (IPC) on detection and development of perioperative venous thromboembolism in brain tumor surgery

Background ObjectiveVenous thromboembolism (VTE), which includes deep vein thrombosis (DVT) and pulmonary embolism (PE), is a common complication in craniotomy patients and is associated with increased morbidity and mortality. The duration of surgery is a known risk factor. Other factors such as positioning and tumor entity have hardly been investigated or are controversial.In two pilot studies, the determination of plasma D-dimer concentration led to a high detection rate of DVT, while the use of intermittent pneumatic venous compression (IPC) drastically reduced the incidence of VTE. In the present study we investigated the efficacy of the two approaches, either alone or in combination, in a large patient cohort.Methods1759 patients who underwent elective craniotomy between 2009 and 2023 were retrospectively analyzed. The staggered use of D-dimer determination and intraoperative use of IPC resulted in 3 groups: Group 1: no procedure; Group 2: D-dimer evaluation; Group 3: IPC and D-dimer evaluation. If the D-dimer level was ≥ 2 mg/l (Fibrinogen equivalent units; FEU), venous ultrasound was performed. Age, gender, tumor entity, duration and extent of surgery, patient positioning, type of VTE were also recorded and analyzed.ResultsThe introduction of postoperative D-dimer evaluation increased the rate of detection of thrombosis from 1.7% in group 1 to 22.6% in group 2. The addition of IPC reduced the rate of thrombosis to 4.4%. Age, gender and patient positioning did not affect the rate of VTE. We were able to confirm the duration of surgery as an individual risk factor and showed that WHO grade 4 tumors and metastasis have an increased VTE risk.ConclusionsIf D-Dimer levels are not analyzed routinely about 20% of craniotomy patients suffer from a clinically silent thrombosis. Each with the risk of fate PE. Intraoperative use of IPC during craniotomy dramatically reduces the risk of VTE.

Diagnostic Accuracy of Tumor Imprint Cytology for Ovarian Cancer

Background: Due to the lack of ovarian cancer screening tools, the combination of clinical findings, serum tumor markers, imaging, and tumor imprint cytology have been studied with the intention of treating the disease. Thus the aim is to determine the diagnostic accuracy of tumor cytology. Methods: Cross-sectional prospective study carried out on surgically treated ovarian tumors. The participants enrolled from the routine surgery list of the Gynecology Department from August 2022 to June 2024. An intraoperative tumor cytology test was done and compared with the final histopathology report. Descriptive statistics were used and diagnostic accuracy was calculated. Results: Out of 82 ovarian tumor surgeries one-third were histopathologically proven malignant tumors and the rest were benign conditions. The most common cancers were of epithelial origin followed by stromal and germ cell origin. The diagnostic accuracy of tumor cytology was 80.5% and two-thirds of cancers were correctly detected. Tumor markers could not consistently guide to characterize the tumor's nature. Conclusion: All malignant tumors had elevated tumor markers but were not exclusive. Tumor cytology together with imaging and tumor markers could help decide the extent of surgery.

Minimally invasive versus open surgery for non-functioning pancreatic neuroendocrine tumors: a systematic review and meta-analysis

Minimally invasive versus open surgery for non-functioning pancreatic neuroendocrine tumors: a systematic review and meta-analysis

Prophylactic cytoreductive surgery and hyperthermic intraperitoneal chemotherapy for low-grade appendiceal mucinous tumors with early and limited disease after completely removed

The necessity of prophylactic cytoreductive surgery (PCRS) and hyperthermic intraperitoneal chemotherapy (HIPEC) for low-grade appendiceal mucinous neoplasms (LAMN) after complete removal is still controversial. This study aims to determine the role of PCRS + HIPEC and identify optimal strategies for managing these patients. One hundred fifty-nine patients who sought medical advice at Aerospace Center Hospital were retrospectively analyzed from January 2011 to December 2021. All the patients were divided into the PCRS group and the observation group. The data of surgical specimens and pathology was collected, and the effect on recurrence-free survival (RFS) was analyzed. Of these 159 patients, 88 were in the PCRS group, and 71 were in the observation group. The median follow-up time was 38 months. Seven patients recurred and developed into pseudomyxoma peritonei, 1 in the PCRS group and 6 in the observation group. The analysis of RFS showed that patients who underwent PCRS (P = .01) and HIPEC (P = .01) had better survival. After multivariate analysis, the surgical specimen accompanied by disseminated peritoneal adenocarcinoma was identified as an independent prognostic factor for RFS. In the study of surgical resection content, patients with greater omentum (P = .01) and bilateral fallopian tubes and ovaries of women (P = .002) resection had a more prolonged RFS with statistical significance. The research indicated that PCRS + HIPEC could prevent recurrence. Therefore this treatment were necessary for LAMN after complete removal in patients with a high risk of recurrence. However, it was not proven to be an independent factor for RFS, and a multicenter, prospective, randomized trial was need to definitively address the role of PCRS + HIPEC for LAMN after complete resection.

Advances in the Prevention and Nursing Care of Venous Thromboembolism After Bone Tumor Surgery

Advances in the Prevention and Nursing Care of Venous Thromboembolism After Bone Tumor Surgery

Current Management of Locally Recurrent Rectal Cancer

Locally recurrent rectal cancer (LRRC), which occurs in 6–12% of patients previously treated with surgery, with or without pre-operative chemoradiation therapy, represents a complex and heterogeneous disease profoundly affecting the patient’s quality of life (QoL) and long-term survival. Its management usually requires a multidisciplinary approach, to evaluate the several aspects of a LRRC, such as resectability or the best approach to reduce symptoms. Surgical treatment is more complex and usually needs high-volume centers to obtain a higher rate of radical (R0) resections and to reduce the rate of postoperative complications. Multiple factors related to the patient, to the primary tumor, and to the surgery for the primary tumor contribute to the development of local recurrence. Accurate pre-treatment staging of the recurrence is essential, and several classification systems are currently used for this purpose. Achieving an R0 resection through radical surgery remains the most critical factor for a favorable oncologic outcome, although both chemotherapy and radiotherapy play a significant role in facilitating this goal. If a R0 resection of a LRRC is not feasible, palliative treatment is mandatory to reduce the LRRC-related symptoms, especially pain, minimizing the effect of the recurrence on the QoL of the patients. The aim of this manuscript is to provide a comprehensive narrative review of the literature regarding the management of LRRC.

Can Preoperative Hounsfield Unit Measurement Help Predict Mechanical Failure in Metastatic Spinal Tumor Surgery?

Can Preoperative Hounsfield Unit Measurement Help Predict Mechanical Failure in Metastatic Spinal Tumor Surgery?

Comparison of 18F-FDG and 68Ga-DOTATATE PET/CT in surgically treated lung carcinoid tumors

Aims: This study aims to compare the efficacy of FDG-18 PET/CT and 68Ga-DOTATATE PET/CT imaging techniques in patients with lung carcinoid tumors, identifying the most appropriate preoperative nuclear medicine technique for diagnosis and staging. Methods: We retrospectively analyzed data from 123 patients who underwent surgery for lung carcinoid tumors at our center between 2009 and 2021. All of the patients were scanned with FDG-18 PET/CT before surgery. In addition to FDG-18 PET/CT, 68Ga-DOTATATE PET/CT scanning was performed in 28 patients: 17 in the preoperative and 11 in the postoperative period. Demographic data, mean higher maximal standard uptake (SUVmax) values of primary mass, lymph nodes and extrathoracic foci, pathologic subtype, and type of surgery were recorded. Compliance with normal distribution for numerical data was assessed using the Shapiro-Wilk test. The Wilcoxon signed-rank test was employed for groups meeting normal distribution to compare continuous numerical variables. The Mann-Whitney U and Kruskal-Wallis tests were used when normal distribution assumptions were unsatisfied. Results: The mean SUVmax value in 68Ga-DOTATATE PET/CT was significantly higher than FDG-18 PET/CT in patients with a lung carcinoid tumor (20 vs 4.4). For 68Ga-DOTATATE PET/CT scan, typical carcinoids had higher mean SUVmax value than atypical carcinoids (26 and 5.6 respectively), and the difference was statistically significant (p=0.002). For that FDG18 PET/CT, on the contrary, the mean SUVmax value was higher in atypical carcinoids than typical carcinoids (5.4 and 3.8, respectively), and the difference was not significant (p=0.126) Conclusion: The SUVmax values from 68Ga-DOTATATE PET/CT and FDG-18 PET/CT in lung carcinoid tumors vary by tumor subtype. 68Ga-DOTATATE PET/CT demonstrates higher SUVmax values in typical carcinoid tumors, indicating its superiority over FDG-18 PET/CT for this subtype. Although 68Ga-DOTATATE PET/CT also shows elevated SUVmax in atypical carcinoid tumors, the difference compared to FDG-18 PET/CT does not reach statistical significance.

Multiphysics simulation of liposome release from hydrogels for cavity filling following patient-specific breast tumor surgery

Several studies have recommended the use of hydrogels for localized targeted delivery of chemotherapeutic drugs following tumor removal surgery. This approach aims to both fill the cavity and prevent cancer recurrence. The use of Multiphysics-based simulation emerges as a valuable strategy for minimizing experimental work, providing detailed insights into how drug release occurs in the tissue, and enabling the optimization of the design.In this study, we introduced a mathematical model, utilizing experimental data, to investigate the transport of liposomes carrying MZ1 from a thermosensitive hydrogel and their impact on the viability of breast cancer cells. The proposed comprehensive model considers not just the transport within the interstitial tissue, represented as a porous medium, but also the uptake by cells and its influence on cell viability, along with the potential lymphatic drainage.The six real patient-specific tumor shapes extracted from MRI scans were used to investigate how the size and form of the tumor can modify the transport pattern. The computational results revealed that the concentration of liposomes in the tissue is significantly influenced by their release from the hydrogel, which proved to be the limiting step. Liposome concentrations of approximately 0.1 % weight were found to be sufficient in ensuring minimal cell survival in the vicinity of the tumor.

Automatic Image Registration Provides Superior Accuracy Compared with Surface Matching in Cranial Navigation.

The precision of neuronavigation systems relies on the correct registration of the patient's position in space and aligning it with radiological 3D imaging data. Registration is usually performed by the acquisition of anatomical landmarks or surface matching based on facial features. Another possibility is automatic image registration using intraoperative imaging. This could provide better accuracy, especially in rotated or prone positions where the other methods may be difficult to perform. The aim of this study was to validate automatic image registration (AIR) using intraoperative cone-beam computed tomography (CBCT) for cranial neurosurgical procedures and compare the registration accuracy to the traditional surface matching (SM) registration method based on preoperative MRI. The preservation of navigation accuracy throughout the surgery was also investigated. Adult patients undergoing intracranial tumor surgery were enrolled after consent. A standard SM registration was performed, and reference points were acquired. An AIR was then performed, and the same reference points were acquired again. Accuracy was calculated based on the referenced and acquired coordinates of the points for each registration method. The reference points were acquired before and after draping and at the end of the procedure to assess the persistency of accuracy. In total, 22 patients were included. The mean accuracy was 6.6 ± 3.1 mm for SM registration and 1.0 ± 0.3 mm for AIR. The AIR was superior to the SM registration (p < 0.0001), with a mean improvement in accuracy of 5.58 mm (3.71-7.44 mm 99% CI). The mean accuracy for the AIR registration pre-drape was 1.0 ± 0.3 mm. The corresponding accuracies post-drape and post-resection were 2.9 ± 4.6 mm and 4.1 ± 4.9 mm, respectively. Although a loss of accuracy was identified between the preoperative and end-of-procedure measurements, there was no statistically significant decline during surgery. AIR for cranial neuronavigation consistently delivered greater accuracy than SM and should be considered the new gold standard for patient registration in cranial neuronavigation. If intraoperative imaging is a limited resource, AIR should be prioritized in rotated or prone position procedures, where the benefits are the greatest.

When Would Minimally Invasive Spinal Surgery Not Be Preferable for Metastatic Spine Disease?

Metastatic spine tumor surgery (MSTS) is an important treatment modality of metastatic spinal disease (MSD). Open spine surgery (OSS) was previously the gold standard of treatment till the early 2010s. However, advancements in MSTS in recent years have led to the advent of minimally invasive spinal surgery (MISS) techniques for the treatment of MSD. The clear benefits of MISS have resulted in a current paradigm shift toward today's gold standard of MISS and early adjuvant radiotherapy in treating MSD patients. Nonetheless, despite improvements in surgical techniques and the rise of literature supporting MISS for MSD, there are still certain situations whereby MISS is not desirable or even suitable. There has also yet to be any literature describing the considerations of not using MISS in MSD in today's clinical context. A narrative review was conducted for this manuscript. All studies related to OSS and MISS in MSTS were included. A total of 54 studies were included in this review. These studies discussed various advantages of MISS for MSD in today's clinical context, including the patient profile, location of vertebrae involved with metastasis requiring treatment, tumor characteristics, as well as equipment availability. This study establishes situations in which MISS can be less applicable despite the advantages it may confer over traditional OSS. MSTS should be individualized, depending on the experience of the surgeon. OSS is a time-tested approach that still holds weight in MSTS and should be readily utilized depending on the clinical situation.

Injectable Polydopamine Nanoparticle-Incorporated Hydrogels for Antiangiogenesis and Stimulating Tumoricidal Immunity to Inhibit Metastasis and Recurrence Postresection.

Surgical resection is still the main means for clinical treatments of breast cancer, but the postoperative immunosuppressive microenvironment and neoangiogenesis of the residual tumors easily lead to tumor metastasis and recurrence, which will further endanger patients' lives. The combination of antiangiogenic therapy and immunotherapy may promote the mutually reinforced cycle of immune reprogramming and vascular normalization to avoid tumor metastasis and recurrence. Herein, we prepared polydopamine nanoparticles for improving tissue adhesion and enriching tumor-associated antigens. This nanoregulator together with regorafenib (REG) was further incorporated into a hydrogel developed from grafting adipic acid dihydrazide onto 2,2,6,6-tetramethylpiperidine-1-oxyl radical oxidized chitin and oxidized hyaluronic acid, which was injectable at the cavity after subcutaneous tumor surgery with good mechanical properties and degradability. The system showed long-term release of REG. After combining with anti-PD-L1, the hydrogel applied to the surgical wound exhibited a reduction in tumor metastasis and recurrence. This effect was achieved by suppressing angiogenesis and enhancing antitumor immunity, characterized by increased levels of effector T lymphocytes and activation of dendritic cells within tumors, spleens, and draining lymph nodes. The injectable hydrogel offers a promising strategy for postoperative management aimed at preventing tumor metastasis and recurrence.

Effects of preoperative neoadjuvant chemotherapy on postoperative delirium in patients with gynecological tumor surgery: an observational study

PurposeTo investigate whether neoadjuvant chemotherapy (NACT) increases the incidence of postoperative delirium (POD) in patients with gynecological tumors undergoing radical hysterectomy.MethodsThis study included 60 patients in the neoadjuvant chemotherapy exposure group and 60 in the non-exposure group. Preoperative cognitive function, the incidence of POD and other physiological parameters were assessed on preoperative day 1 (POD-1), postoperative day 1 (POD1), postoperative day 2 (POD2), and postoperative day 3 (POD3). Additionally, preoperative olfactory function was evaluated using an olfactory detection kit on POD-1. The primary outcome was the incidence of POD within three days after surgery.ResultsThe incidence of POD was 28.33% in the exposed group and 8.33% in the non-exposed group (P = 0.005). Compared to the non-exposed group, the exposed group had a higher rate of cognitive dysfunction (33.33% vs 13.33%; P = 0.010), and a higher rate of olfactory dysfunction (OD) (25.00% vs 10.00%; P = 0.031). A restricted cubic spline analysis revealed a non-linear relationship between olfactory test scores and Montreal Cognitive Assessment Scale (MoCA) scores (P for overall < 0.001, P for nonlinear = 0.001). Logistic regression identified NACT, mild and moderate cognitive dysfunction, OD, and depression as independent risk factors for POD, with all factors showing significant associations (P < 0.05). The area under the curve (AUC) of OD for predicting POD was 0.783 (95%CI 0.656–0.909).ConclusionsThis single-blind observational study suggests that NACT increases the incidence of POD in patients with gynecological tumors undergoing radical hysterectomy. Moreover, the results indicate that preoperative OD may reflect preoperative cognitive dysfunction, and have predictive value for the incidence of POD.

OGC SO12 - Extended Venous Thromboembolism Prophylaxis in Postoperative Oesophagogastric Cancer Patients: A Completed Audit

Abstract Background Venous thromboembolism (VTE) is a life-threatening complication for cancer patients undergoing oesophagogastric surgery. The NICE recommends extending pharmacological VTE prophylaxis to 28 days for patients undergoing major abdominal cancer surgery. While the effectiveness of in-hospital VTE prophylaxis is well-documented, the optimal duration for postoperative VTE prophylaxis remains a subject of debate. This study aims to evaluate the prescribing compliance, effectiveness, and safety of extending VTE prophylaxis to 28 days following hospital discharge (HD), a policy adopted trust-wide since 2014. Method This is an 11-year retrospective study from July 2012-2022, followed by a complete audit cycle conducted between January 2023-2024. The study included all patients who underwent cardio-esophagectomy or gastrectomy for oesophagogastric cancers. Exclusion criteria comprised patients who died during the index admission, those who underwent benign oesophagogastric surgery or local excision of gastrointestinal stromal tumours (GIST), and patients with peripherally inserted central catheter (PICC) line-associated VTE. The primary outcome measured was the prescribing compliance of 28-day VTE prophylaxis post-hospital discharge. Secondary endpoints included the incidence of VTE and bleeding complications necessitating hospital readmission within 28 days of discharge. Results Our study included 582 patients, showing that post-2014, 423 out of 490 patients (86.3%) were discharged with 28-day VTE prophylaxis, compared to 13 out of 92 patients (14.1%) pre-2014, indicating a significant improvement (p&amp;lt;0.05). Only one patient (0.17%) developed VTE more than 28 days post-discharge, and no readmissions due to bleeding complications. In the subsequent audit, 63 out of 68 patients (92.6%) received the prophylaxis, further increasing prescribing compliance. Comparing the two periods, compliance improved from 86.3% to 92.6%, though not statistically significant (p=0.145). Overall, VTE incidence was minimised from 0.17% to 0%, with no readmissions due to bleeding complications. Conclusion Routine extended VTE prophylaxis for 28 days post-hospital discharge should be recommended for cancer patients undergoing oesophagogastric surgery. This practice has significantly improved prescribing compliance rates, reduced post-hospital discharge VTE incidence, and demonstrated safety, with no hospital readmissions due to bleeding complications.

Hyperspectral imaging in neurosurgery: a review of systems, computational methods, and clinical applications.

Accurate identification between pathologic (e.g., tumors) and healthy brain tissue is a critical need in neurosurgery. However, conventional surgical adjuncts have significant limitations toward achieving this goal (e.g., image guidance based on pre-operative imaging becomes inaccurate up to 3cm as surgery proceeds). Hyperspectral imaging (HSI) has emerged as a potential powerful surgical adjunct to enable surgeons to accurately distinguish pathologic from normal tissues. We review HSI techniques in neurosurgery; categorize, explain, and summarize their technical and clinical details; and present some promising directions for future work. We performed a literature search on HSI methods in neurosurgery focusing on their hardware and implementation details; classification, estimation, and band selection methods; publicly available labeled and unlabeled data; image processing and augmented reality visualization systems; and clinical study conclusions. We present a detailed review of HSI results in neurosurgery with a discussion of over 25 imaging systems, 45 clinical studies, and 60 computational methods. We first provide a short overview of HSI and the main branches of neurosurgery. Then, we describe in detail the imaging systems, computational methods, and clinical results for HSI using reflectance or fluorescence. Clinical implementations of HSI yield promising results in estimating perfusion and mapping brain function, classifying tumors and healthy tissues (e.g., in fluorescence-guided tumor surgery, detecting infiltrating margins not visible with conventional systems), and detecting epileptogenic regions. Finally, we discuss the advantages and disadvantages of HSI approaches and interesting research directions as a means to encourage future development. We describe a number of HSI applications across every major branch of neurosurgery. We believe these results demonstrate the potential of HSI as a powerful neurosurgical adjunct as more work continues to enable rapid acquisition with smaller footprints, greater spectral and spatial resolutions, and improved detection.

Foundation models for fast, label-free detection of glioma infiltration.

A critical challenge in glioma treatment is detecting tumour infiltration during surgery to achieve safe maximal resection1-3. Unfortunately, safely resectable residual tumour is found in the majority of patients with glioma after surgery, causing early recurrence and decreased survival4-6. Here we present FastGlioma, a visual foundation model for fast (<10 s) and accurate detection of glioma infiltration in fresh, unprocessed surgical tissue. FastGlioma was pretrained using large-scale self-supervision (around 4 million images) on rapid, label-free optical microscopy, and fine-tuned to output a normalized score that indicates the degree of tumour infiltration within whole-slide optical images. In a prospective, multicentre, international testing cohort of patients with diffuse glioma (n = 220), FastGlioma was able to detect and quantify the degree of tumour infiltration with an average area under the receiver operating characteristic curve of 92.1 ± 0.9%. FastGlioma outperformed image-guided and fluorescence-guided adjuncts for detecting tumour infiltration during surgery by a wide margin in a head-to-head, prospective study (n = 129). The performance of FastGlioma remained high across diverse patient demographics, medical centres and diffuse glioma molecular subtypes as defined by the World Health Organization. FastGlioma shows zero-shot generalization to other adult and paediatric brain tumour diagnoses, demonstrating the potential for our foundation model to be used as a general-purpose adjunct for guiding brain tumour surgeries. These findings represent the transformative potential of medical foundation models to unlock the role of artificial intelligence in the care of patients with cancer.

Nasoseptal Flap Reuse for Reconstruction in Revision Endoscopic Skull Base Surgery

Abstract Introduction The nasoseptal flap (NSF) is the “workhorse” for reconstruction in endoscopic skull base surgery. It creates a reliable, watertight closure to prevent cerebrospinal fluid (CSF) leaks. The flap's robust vascular pedicle allows for it to be taken down and reused in subsequent reconstructive surgeries if needed. Previous literature has demonstrated successful reuse of the flap up to 24 months after initial surgery. The present study analyzes a cohort of patients with NSFs that were reused up to 12 years following the initial surgery. Methods A retrospective analysis was performed on patients who underwent revision endoscopic skull base surgery for recurrent tumor at a tertiary academic skull base center. Patient demographics, tumor type, length of time between first and second surgery, and outcomes were evaluated. Results A total of 15 patients were identified between 2012 and 2022 who underwent revision endoscopic skull base surgery with previous NSFs that were reused for reconstruction. The time interval between surgeries ranged from 9 months to 12 years. There were 12 pituitary adenomas, 2 Rathke's cleft cysts, and 1 craniopharyngioma. One patient developed a CSF leak 5 days postoperatively, which was repaired with an abdominal fat graft, synthetic dural substitute, and a middle turbinate flap. The operative technique for the take-down and reuse of the NSF will be demonstrated. Conclusion The NSF can be taken down and reused in revision endoscopic skull base cases, even several years after the initial surgery, with excellent coverage of the defect and low incidence of postoperative CSF leak.

NIMG-18. RAPID INTRAOPERATIVE DIFFERENTIATION OF PRIMARY CNS LYMPHOMA FROM COMMON CNS TUMORS USING STIMULATED RAMAN HISTOLOGY AND DEEP LEARNING

Abstract Accurate intraoperative diagnosis is crucial for decision-making during tumor surgery and stereotactic-guided biopsies. Differentiating between primary CNS lymphoma (PCNSL) and non-PCNSL entities such as gliomas and metastatic cancers presents significant challenges due to overlapping histomorphological features, time constraints, and the differing further therapy strategies required. Our study utilized deep learning and stimulated Raman histology (SRH) to address this challenge. We imaged unprocessed, label-free tissue samples intraoperatively using a portable Raman scattering microscope, generating virtual H&amp;E-like images within less than five minutes. Our training set comprised over 54,000 SRH tumor patch images, including various brain tumors. We developed a deep learning pipeline utilizing a ResNet-50 feature extractor and the BYOL (Bootstrap Your Own Latent) self-supervised learning algorithm. We trained a binary linear classifier to distinguish between PCNSL and non-PCNSL entities. Data were collected from New York University (NYU), the University of Cologne (UKK), and the University of Michigan (UM), including 263 SRH whole slide images from 177 patients sourced from surgical resections and stereotactic-guided biopsies. In an international bicentric test cohort (N = 100, NYU and UKK), the model achieved an overall accuracy of 91.1%, an area under the receiver operating characteristic curve (AUROC) of 0.983, a sensitivity of 89.3%, and a specificity of 100% compared to the final formalin-fixed, paraffin-embedded-based diagnosis. Our algorithm can accurately extract typical histomorphological features to help differentiate between PCNSL and non-PCNSL tumors. These results underscore the potential of deep learning models to enhance the accuracy of intraoperative diagnoses for PCNSL and non-PCNSL entities, thereby improving intraoperative decision-making and further treatment strategies by supporting neuropathologists and surgeons.