Distinguishing Tumor Research Articles

Discussion on pathological diagnosis of vascular diseases based on ISSVA classification

The diagnosis of vascular diseases is not only about distinguishing neoplastic or non-neoplastic lesions, but also focusing more on emphasizing the essence of the disease, namely the presence or absence of endothelial cell proliferation, and further to distinguishing true hemangioma tumors or vascular malformation. This article is based on the International Society for the Study of Vascular Anomalies (ISSVA) classification, which is widely used in clinical practice, and discusses the related pathological diagnosis issues of vascular diseases.

Quantification of oxygen consumption in head and neck cancer using fluorescent sensor foil technology.

Head and neck squamous cell carcinoma (HNSCC) patients suffer from frequent local recurrences that negatively impact on prognosis. Hence, distinguishing tumor and normal tissue is of clinical importance as it may improve the detection of residual tumor tissue in surgical resection margins and during imaging-based surgery planning. Differences in O2 consumption (OC) can be used to this aim, as they provide options for improved surgical, image-guided approaches. In the present study, the potential of a fluorescent sensor foil-based technology to quantify OC in HNSCC was evaluated in an in vitro 3D model and in situ in patients. In vitro measurements of OC using hypopharyngeal and esophageal cell lines allowed a specific detection of tumor cell spheroids embedded together with cancer-associated fibroblasts in type I collagen extracellular matrix down to a diameter of 440 µm. Pre-surgery in situ measurements were conducted with a handheld recording device and sensor foils with an oxygen permeable membrane and immobilized O2-reactive fluorescent dyes. Lateral tongue carcinoma and carcinoma of the floor of the mouth were chosen for analysis owing to their facilitated accessibility. OC was evaluated over a time span of 60 seconds and was significantly higher in tumor tissue compared to healthy mucosa in the vicinity of the tumor. Hence, OC quantification using fluorescent sensor foil-based technology is a relevant parameter for the differentiation of tumor tissue of the head and neck region and may support surgery planning.

Prediction of Benefit From Adjuvant Pertuzumab by 80-Gene Signature in the APHINITY (BIG 4-11) Trial.

At the primary analysis, the APHINITY trial reported a statistically significant but modest benefit of adding pertuzumab to standard adjuvant chemotherapy plus trastuzumab in patients with histologically confirmed human epidermal growth factor receptor 2 (HER2)-positive early-stage breast cancer. This study evaluated whether the 80-gene molecular subtyping signature (80-GS) could identify patients within the APHINITY population who derive the most benefit from dual anti-HER2 therapy. In a nested case-control study design of 1,023 patients (matched event to control ratio of 3:1), the 80-GS classified breast tumors into functional luminal type, HER2 type, or basal type. Additionally, 80-GS distinguished tumor subtypes that exhibited a single-dominant functional pathway versus tumors with multiple activated pathways. The primary end point was invasive disease-free survival (IDFS). Hazard ratios (HRs) were evaluated by Cox regression. After excluding patients without appropriate consent and those with missing data, 964 patients were included. The 80-GS classified 50% (n = 479) of tumors as luminal type, 28% (n = 275) as HER2 type, and 22% (n = 209) as basal type. Most luminal-type tumors (86%) displayed a single-activated pathway, whereas 49% of HER2-type and 42% of basal-type tumors were dual activated. There was no significant difference in IDFS among different conventional 80-GS subtypes (single- and dual-activated subtypes combined). However, basal single-subtype tumors were significantly more likely to have an IDFS event (hazard ratio, 1.69 [95% CI, 1.12 to 2.54]) compared with other subtypes. HER2 single-subtype tumors displayed a trend toward greater beneficial effect on the addition of pertuzumab (hazard ratio, 0.56 [95% CI, 0.27 to 1.16]) compared with all other subtypes. The 80-GS identified subgroups of histologically confirmed HER2-positive tumors with distinct biological characteristics. Basal single-subtype tumors exhibit an inferior prognosis compared with other subgroups and may be candidates for additional therapeutic strategies. Preliminary results suggest patients with HER2-positive, genomically HER2 single-subtype tumors may particularly benefit from added pertuzumab, which warrants further investigation.

Abstract B013: Development of oropharyngeal tumor organoids for disease modeling and high throughput drug screening

Abstract Head and neck squamous cell carcinoma (HNSCC) is the sixth most common cancer worldwide with approximately 550,000 new cases diagnosed each year. The oropharynx is one of the most common sites of occurrence, and disease in this region most often presents in the palatine tonsil, followed by the base of tongue and soft palate. Human papilloma virus (HPV) is a common cause of HNSCC, giving rise to a distinct subset of tumors with a different clinical presentation, disease course, and response to treatment. In order to further characterize HPV+ HNSCC tumors and their therapy response, we have leveraged our high throughput 3D tumor organoid screening platform. Our approach entails seeding tumor cells in extracellular matrix gels deposited around the perimeter of tissue culture wells in ring-like structures, with screening results available within a week from surgery (Phan et al, 2019; Al Shihabi et al, 2022; Nguyen et al, 2022; Al Shihabi et al 2023). We particularly focused on tonsil tumors, and develop organoids from both HNSCC and contralateral normal tonsil tissue. Thus far, we have procured tonsil samples from 31 patients to develop organoids, perform high throughput drug screening, and immunohistochemical characterization including H&E staining, Ki67, p16 and p53 immunohistochemical staining. We have leveraged dynamic optical contrast imaging (DOCI), a label-free tissue autofluorescence imaging method that is being investigated clinically (Kim et al, 2017), to assist in distinguishing tumor organoids from normal tissue organoids. Lastly, given the importance of radiation as a treatment modality for HPV+ HNSCC, we have also optimized a chemoradiation pipeline to test possible combinatorial effects. Taken together, our results demonstrate that 3D organoids of HNSCC generated from patient samples can successfully preserve native tissue characteristics and can be used for high throughput drug screening and radiation de-escalation testing, affording critical insight into the management of HNSCC in patients. Citation Format: Krishna K. Bommakanti, Luda Lin, Yazeed Alhiyari, Andreanne Sannajust, Brandon Mo, Lauran Evans, Peyton Tebon, Maie St. John, Alice Soragni. Development of oropharyngeal tumor organoids for disease modeling and high throughput drug screening [abstract]. In: Proceedings of the AACR Special Conference in Cancer Research: Translating Cancer Evolution and Data Science: The Next Frontier; 2023 Dec 3-6; Boston, Massachusetts. Philadelphia (PA): AACR; Cancer Res 2024;84(3 Suppl_2):Abstract nr B013.

FLI-1 is expressed in a wide variety of hematolymphoid neoplasms: a special concern in the differential diagnosis

Friend Leukemia Virus Integration 1 (FLI-1) is a member of E26 transformation-specific family of transcription factors that participates in hematopoietic and vascular endothelial cell development. Immunohistochemical detection of FLI-1 has been widely used to diagnose vascular tumors or, more evidently, Ewing’s sarcoma. However, the expression pattern of FLI-1 in hematolymphoid neoplasms remains unclear. Therefore, in this study, we aimed to investigate the expression of FLI-1 in these tumors, focusing on high-grade lesions, which presents a diagnostic challenge by mimicking Ewing’s sarcoma. We evaluated the expression FLI-1 in various types of lymphoid and plasmacytic tumors, including 27 plasmablastic lymphomas, 229 diffuse large B-cell lymphomas, 22 precursor T- or B-lymphoblastic lymphomas, 24 angioimmunoblastic-type nodal T-follicular helper cell lymphomas, 52 peripheral T-cell lymphomas, NOS, 18 Burkitt lymphomas, 18 non-gastric lymphomas of mucosa-associated lymphoid tissue, 38 chronic lymphocytic leukemia/small lymphocytic lymphomas, 15 mantle cell lymphomas, 23 gastric MALT lymphomas, 50 plasma cell myelomas, and 38 follicular lymphomas. We calculated the H-scores of FLI-1 immunostaining, ranging from 0 to 200, and used the scores to analyze the clinicopathological significance of FLI-1 statistically. FLI-1 was expressed to varying degrees in all types of hematological tumors. FLI-1 expression was detected in 84.1% of patients (466/554). FLI-1 was highly expressed in precursor T- or B-lymphoblastic lymphomas. Follicular lymphomas exhibited low FLI-1 expression. In plasmablastic lymphoma, 85.2% of the patients were focally positive for FLI-1. FLI-1 expression did not correlate with clinicopathological variables, such as demographic data or disease stage, in patients with plasmablastic lymphoma and diffuse large B-cell lymphoma. However, FLI-1 overexpression was associated with poorer overall survival in patients with plasmablastic lymphoma. This study demonstrates that FLI-1 is expressed in various hematolymphoid neoplasms. FLI-1 expression can lead to diagnostic confusion, especially in small blue round cell tumors, such as lymphoblastic lymphoma, plasmablastic lymphoma, and plasma cell myeloma, when distinguishing tumors positive for CD99 and CD56 without CD3, CD20, or CD45. Our findings also suggested the possibility of FLI-1 as a potential prognostic biomarker for plasmablastic lymphoma.

Abstract C008: MUC16 expression in pancreatic ductal adenocarcinoma patient samples after neoadjuvant chemotherapy

Abstract Purpose Pancreatic ductal adenocarcinoma (PDAC) is a deadly disease with low five-year survival rates. Currently, surgical resection is the only potentially curative treatment available to patients. However, few patients are eligible for surgery without receiving prior therapy due to disease progression. Of those that receive upfront surgery, up to 60% develop recurrent disease within 5 years. Therefore, there is a need to increase the rate of complete surgical resections and the number of eligible patients for surgery. In the last decade, neoadjuvant therapy or treatment prior to surgery has increased for patients diagnosed with borderline resectable and locally advanced PDAC. Often, this therapy consists of chemotherapy alone or chemoradiotherapy treatment intended to downstage the tumor and reduce micrometastatic disease. Fluorescence-guided surgery (FGS) is a potential tool to increase the rate of complete surgical resections by specifically targeting a protein biomarker upregulated in the tumor with a fluorescent dye conjugated probe. Previous in vitro and in vivo research from our lab has indicated that the transmembrane mucin, MUC16, is a promising target for FGS in PDAC. However, protein expression has the potential to deviate after chemotherapy treatment. Therefore, this research aimed to determine the effect of neoadjuvant chemotherapy treatment (NCT) on the FGS-targetable biomarker, MUC16, in PDAC patient specimens. Experimental Procedures Deidentified PDAC and matched adjacent patient specimens receiving the NCT regimen or surgery-first treatment were obtained from UNMC Pathology and the UNMC Paraffin Tissue Bank. Deidentified normal pancreas specimens were received from the Normal Organ Recovery program, UNMC. MUC16 expression was investigated in each specimen. The presence of PDAC tumor and MUC16 stain expression patterns were confirmed by a specimen-blinded, board-certified pathologist. The expression of MUC16 between the NCT, surgery-first, and normal pancreas specimens was compared via statistical analysis. Results MUC16 was expressed in both the NCT and surgery-first specimens. There was a significant difference in MUC16 expression between the NCT specimens and both the normal pancreas (P = 0.0450) and surgery-first (P = 0.0494) specimens. In matched adjacent tissue, MUC16 expression was found to be higher in the NCT (P = 0.0003) and surgery-first (P = 0.1446) tumor specimens allowing for distinct tumor margins; however, this difference was not significant in the surgery-first specimens. Conclusions MUC16 expression retention in NCT PDAC specimens may indicate that MUC16 is a stable biomarker after chemotherapy treatment. Distinct MUC16 expression between tumor and matched adjacent specimens indicates the ability to achieve distinguishable tumor margins allowing for complete tumor resection. Therefore, MUC16 remains targetable for FGS in chemotherapy-treated patients. This has the potential to broaden the benefiting patient population and increase the number of patients that may undergo a complete surgical resection. Citation Format: Kathryn M. Muilenburg, Evie G. Ehrhorn, Madeline T. Olson, Carly C. Isder, Kelsey A. Klute, Geoffrey A. Talmon, Mark A. Carlson, Quan P. Ly, Aaron M. Mohs. MUC16 expression in pancreatic ductal adenocarcinoma patient samples after neoadjuvant chemotherapy [abstract]. In: Proceedings of the AACR Special Conference in Cancer Research: Pancreatic Cancer; 2023 Sep 27-30; Boston, Massachusetts. Philadelphia (PA): AACR; Cancer Res 2024;84(2 Suppl):Abstract nr C008.

YRNA and tRNA fragments can differentiate benign from malignant canine mammary gland tumors

Mammary gland tumors (MGT) are the most common tumors in sexually intact female dogs. The functional regulation of miRNAs, a type of noncoding RNAs (ncRNAs), in canine MGT has been extensively investigated. However, the expression of other ncRNAs, such as YRNAs and transfer RNA–derived fragments (tRFs) in canine MGT is unknown. We investigated ncRNAs other than miRNAs from our small RNA project (PRJNA716131) in different canine MGT histologic subtypes. This study included benign tumors (benign mixed tumor, complex adenoma) and malignant tumors (carcinoma in benign tumor and carcinoma with metastasis) samples. Aberrantly expressed ncRNAs were examined by comparisons among MGT subtypes. The relative expression trends were validated in canine MGT tissues, plasma, extracellular vesicles, and MGT cell lines using quantitative reverse transcription PCR. Three aberrantly expressed ncRNAs were identified by comparisons among MGT subtypes. YRNA and tRNA-Gly-GCC distinguished benign mixed tumor from other MGT histologic subtypes, while tRNA-Val differentiated complex adenoma, carcinoma in benign tumors, and carcinoma with metastasis. The ROC curve of the three ncRNAs showed they might be potential biomarkers to discriminate malignant from benign MGT. YRNA and tRFs expression levels were decreased in metastatic compared with primary canine MGT cell lines. To the best of our knowledge, this is the first investigation of YRNA and tRFs in canine MGT. The three identified ncRNAs may be biomarkers for differentiating MGT histologic subtypes. Suggested Reviewers: Powered by Editorial Manager® and ProduXion Manager® from Aries Systems Corporatio

NIMG-30. DEEP LEARNING SEGMENTATION OF INFILTRATIVE AND ENHANCING CELLULAR TUMOR IN POST-TREATMENT GLIOBLASTOMA PATIENTS WITH MULTI-SHELL DIFFUSION MRI

Abstract Multimodal MRI is used for the evaluation of tumor burden in glioblastoma (GBM) patients after surgical resection, radiation, and chemotherapy. However, it can be challenging to identify non-enhancing infiltrative tumor and differentiate recurrent enhancing tumor from post-treatment changes. Restriction spectrum imaging (RSI), an advanced multishell diffusion technique, has shown promise in distinguishing cellular tumor from edema and enhancing post-treatment changes including pseudoprogression and radiation necrosis. We developed a convolutional network (nnUNet) that incorporates traditional multimodal MRI with DSC perfusion and RSI to detect and quantify enhancing and nonenhancing cellular tumor in post-treatment glioblastoma patients. This cellular tumor segmentation network was evaluated in a cohort of 186 post-treatment GBM patients (55.5 +/- 13.7 years, 158 male) with 243 timepoints using 5-fold cross validation. Cellular tumor volumes were then used to predict overall survival (OS) and progression free survival (PFS) in 94 patients with imaging within 3 months of initial surgery. Accuracy for segmenting cellular tumor as measured by the Dice-similarity coefficient was 0.77 (IQR 0.55-0.87) and the AUC of the ROC curve for detecting residual/recurrent tumor compared to post-treatment changes was 0.86. The predicted cellular tumor volumes were an independent hazard over age, sex, GTR status and conventional enhancing residual tumor for both OS (p=0.0003) and PFS (p=0.0004). The tumor segmentation network was validated in an external dataset of 22 GBM patients (55 timepoints) with a median Dice of 0.73 and AUC of 0.91. Survival prediction was validated in a different external dataset of 70 post-treatment GBM patients, with cellular tumor volumes being an independent predictor of OS (p=0.03). In conclusion, deep learning models incorporating advanced imaging can accurately segment enhancing and nonenhancing cellular tumor, which can help distinguish recurrent/residual tumor from post-treatment changes and predict OS and PFS.

INNV-24. MACHINE LEARNING-BASED SPECTROSCOPIC TISSUE DIFFERENTIATION IN FLUORESCENCE-GUIDED NEUROSURGERY

Abstract Maximal resection of malignant gliomas is hindered by difficulty in distinguishing tumor margins. Fluorescence-guided resection with 5-ALA assists in reaching this goal. Previously, we characterized five fluorophore emissions that accurately represent any spectrum measured from human brain tumor biopsies with a wide-field hyperspectral device. In this study, the effectiveness of these five spectra was explored for different tumor classification tasks in 891 hyperspectral widefield measurements of 184 patients harboring low- (n=30) and high-grade gliomas (n=115), non-glial primary brain tumors (n=19), radiation necrosis (n=2), miscellaneous (n=10) and metastases (n=8), which corresponds to up to 15000 spectra for a given test. The statistical differences in fluorophore abundances between classes were determined and visualized using dimensionality reduction techniques, including principal component analysis (PCA) and t-distributed stochastic neighbor embedding (t-SNE). Three machine-learning models were trained to classify tumor type (12 classes), grade (3 classes), and glioma margins (3 classes). Five algorithms were tested with varying hyperparameters for each classification task. We explored whether PCA projection onto five different axes than fluorophore abundances can provide more information for visualization and classification. The abundances of the five a priori fluorophore spectra matched or outperformed the five optimal PCA components for all classification tasks. These five axes capture 96-99% of the variance in the dataset. Using random forests and multilayer perceptrons, the classifiers achieved average test accuracies of 74-82%, 79%, and 81%, respectively. All five fluorophore abundances varied between tumor margin type as well as between grades (p < 0.01). For tissue type, at least four of five fluorophore abundances were found to be significantly different (p < 0.01) between all classes. These results demonstrate the differing contribution of fluorophores in different tissue classes tissue, as well as the five fluorophores value as potential optical biomarkers, opening new opportunities for intraoperative classification systems in fluorescence-guided neurosurgery.

One Stone, Three Birds: Multifunctional Nanodots as "Pilot Light" for Guiding Surgery, Enhanced Radiotherapy, and Brachytherapy of Tumors.

Surgery, radiotherapy (RT), and brachytherapy are crucial treatments for localized deep tumors. However, imprecise tumor location often leads to issues such as positive surgical margins, extended radiotherapy target volumes, and radiation damage to healthy tissues. Reducing side effects in healthy tissue and enhancing RT efficacy are critical challenges. To address these issues, we developed a multifunctional theranostic platform using Au/Ag nanodots (Au/AgNDs) that act as a "pilot light" for real-time guided surgery, high-efficiency RT, and brachytherapy, achieving a strategy of killing three birds with one stone. First, dual-mode imaging of Au/AgNDs enabled precision RT, minimizing damage to adjacent normal tissue during X-ray irradiation. Au/AgNDs enhanced ionizing radiation energy deposition, increased intracellular reactive oxygen species (ROS) generation, regulated the cell cycle, promoted DNA damage formation, and inhibited DNA repair in tumor cells, significantly improving RT efficacy. Second, in brachytherapy, precise guidance provided by dual-mode imaging addressed challenges related to non-visualization of existing interstitial brachytherapy and multiple adjustments of insertion needle positions. Meanwhile, the effect of brachytherapy was improved. Third, the excellent fluorescence imaging of Au/AgNDs accurately distinguished tumors from normal tissue, facilitating their use as a powerful tool for assisting surgeons during tumor resection. Taken together, our multifunctional theranostic platform offers real-time guidance for surgery and high-efficiency RT, and improves brachytherapy precision, providing a novel strategy and vision for the clinical diagnosis and treatment of cancer.

Fluorescence lifetime of injected indocyanine green as a universal marker of solid tumours in patients.

The surgical resection of solid tumours can be enhanced by fluorescence-guided imaging. However, variable tumour uptake and incomplete clearance of fluorescent dyes reduces the accuracy of distinguishing tumour from normal tissue via conventional fluorescence intensity-based imaging. Here we show that, after systemic injection of the near-infrared dye indocyanine green in patients with various types of solid tumour, the fluorescence lifetime (FLT) of tumour tissue is longer than the FLT of non-cancerous tissue. This tumour-specific shift in FLT can be used to distinguish tumours from normal tissue with an accuracy of over 97% across tumour types, and can be visualized at the cellular level using microscopy and in larger specimens through wide-field imaging. Unlike fluorescence intensity, which depends on imaging-system parameters, tissue depth and the amount of dye taken up by tumours, FLT is a photophysical property that is largely independent of these factors. FLT imaging with indocyanine green may improve the accuracy of cancer surgeries.

P11.07.A PERFUSION-MRI BASED DIFFERENTIATION BETWEEN EARLY TUMOR PROGRESSION AND PSEUDOPROGRESSION IN GLIOBLASTOMA AND ITS USE IN CLINICAL PRACTICE

Abstract BACKGROUND Post-gadolinium MRI is considered the mainstay of radiological monitoring of treatment effects in patients with a glioblastoma. However, the radiotherapy(RT)-induced abnormality, pseudoprogression, is difficult to distinguish from real tumor growth. Perfusion MRI (pMRI), including Arterial Spin Labeling (ASL) and Dynamic Susceptibility Contrast (DSC), may help in this differentiation. Evaluation of RT effects is usually discussed during multidisciplinary team meetings (MDTM), after which a decision regarding (dis)continuation of anti-tumor treatment is made. It is currently unknown whether pMRI impacts such treatment decisions. This study evaluated how accurate pMRI and the MDTM are in distinguishing tumor progression from pseudoprogression in glioblastoma. MATERIAL AND METHODS Population and outcome: Glioblastoma patients who received postoperative RT were retrospectively included from 2015 to 2020. ASL, DSC, FLAIR and post-gadolinium 3D-T1 at 3 months post-RT were visually scored for probability of radiological tumor progression. These radiological scores, as well as the outcome of MDTM, were compared to the final outcome (i.e. tumor progression or pseudoprogression) set at 9 months after RT initiation. Statistical analysis: Diagnostic accuracy of pMRI, conventional MRI and MDTM was assessed by 2x2 tables, sensitivity and specificity (including AUC values). In case the MDTM was inconclusive, subsequent analysis was used to determine whether pMRI has additional value. A multivariable logistic regression model was constructed to assess which clinical parameters (age, surgery type, IDH mutation status, MGMT methylation status, KPS score, chemotherapy and pMRI evaluation) were associated with the outcome of the MDTM assessment. RESULTS Sixty-five patients with a glioblastoma met the inclusion criteria (40 male, mean age 60+/-13 years, median KPS of 90). For all METHODS , the sensitivity for detecting true disease progression was poor to moderate (32-62%). Although sensitivity was higher when adding pMRI to conventional imaging (56% versus 32%), specificity was lower (80% versus 93%). AUC values were comparable (range 0.63-0.74), but highest for MDTM evaluation (0.74). For the 26% of patients where MDTM was inconclusive, pMRI shows increased sensitivity compared to conventional MRI in detecting tumor progression (respectively 36% versus 0%). The multivariable regression analysis showed significant associations for KPS score (OR = 0.84 (95% CI: 0.77-0.91)) and radiological assessment based on pMRI with MDTM outcome (OR = 0.09 (95% CI: 0.02-0.52)). CONCLUSION This study shows that perfusion MRI has added value for detection of tumor progression within the MDTM assessment in clinical decision-making for glioblastomas, especially when the MDTM outcome is uncertain. For this decision, the role of perfusion MRI and the KPS score are most important.

More than bad luck: Cancer and aging are linked to replication-driven changes to the epigenome.

Aging is a leading risk factor for cancer. While it is proposed that age-related accumulation of somatic mutations drives this relationship, it is likely not the full story. We show that aging and cancer share a common epigenetic replication signature, which we modeled using DNA methylation from extensively passaged immortalized human cells in vitro and tested on clinical tissues. This signature, termed CellDRIFT, increased with age across multiple tissues, distinguished tumor from normal tissue, was escalated in normal breast tissue from cancer patients, and was transiently reset upon reprogramming. In addition, within-person tissue differences were correlated with predicted lifetime tissue-specific stem cell divisions and tissue-specific cancer risk. Our findings suggest that age-related replication may drive epigenetic changes in cells and could push them toward a more tumorigenic state.

Development of prediction model to estimate future risk of ovarian lesions: A multi-center retrospective study

BackgroundTo develop the preoperative prediction of ovarian lesions using regression-based statistics analyses and machine learning methods based on multiple serological biomarkers in China. Methods1137 patients with ovarian lesions in Zhujiang Hospital and 518 patients in others hospital in China were randomly assigned to training, test and external validation cohorts. Five machine learning classifiers, including Random Forest (RF), Extreme Gradient Boosting (XGB), Support Vector Classifier (SVC), K-nearest Neighbor (KN), Multi-Layer Perceptron (MLP) and the Lasso-Logistics prediction model (LLRM) were used to derive diagnostic information from 23 predictors. ResultsThe RF model had a high diagnostic value (AUC = 0.968) in predicting benign and malignant ovarian disease. Age and MLR were also potential diagnostic indicators for predicting ovarian disease except tumor indicators. The RF model well distinguished borderline ovarian tumors (AUC = 0.742). The RFM had a high predictive power to identify ovarian serous adenocarcinoma (AUC = 0.943) and ovarian endometriosis cysts (AUC = 0.914). ConclusionsThe RF models can effectively predict adnexal lesions, promising to be adjuncts to the preoperative prediction of ovarian cancer.

Absolute monocyte count has a diagnostic role in distinguishing tumor marker-negative TGCT from benign testicular tumor via CCL2 regulation.

Clinically, for testicular tumor patients with negative tumor markers, how to distinguish the malignant from the benign is a difficult problem. This study aimed to assess the clinical significance of the absolute monocyte count (AMC) in differential diagnosis of testicular germ cell tumor with stage S0 (TGCTS0) and benign testicular tumor. In this retrospective single-center study, a total of 90 patients newly diagnosed with benign testicular tumor or TGCTS0 were reviewed. All patients received surgical intervention as the primary treatment method. AMC and other clinicopathological parameters were analyzed. Receiver operating characteristic (ROC) curves were constructed to assess the diagnostic power of investigated parameters, and to determine the optimal cutoff values. Kaplan-Meier curve analysis was used to study the survival of patients with TGCTS0. qRT-PCR and immunohistochemistry (IHC) were performed to examine the expression of C-C motif chemokine ligand 2 (CCL2) mRNA and protein respectively. Differential gene expression and functional enrichment analysis were performed using Gene Expression Omnibus and the Cancer Genome Atlas databases. The mean preoperative AMC in patients with TGCTS0 was significantly higher than that in patients with benign testicular tumor (P = .020). AMC > 0.485*10^9/L was identified to be associated with the presence of TGCTS0 (hazard ratio [HR] = 3.074, P = .026), and patients with higher AMC level had worse progression free survival (PFS) (P = .047). Furthermore, AMC combined with lactate dehydrogenase (LDH) achieved a better diagnostic efficacy for TGCTS0 (area under curve [AUC] = 0.695). Tumor-associated macrophages (TAMs) signature gene CCL2 was highly expressed in TGCT compared with normal testicular tissue. Functional enrichment analysis showed that CCL2 is closely involved in the Extracellular Matrix Organization pathway and positively correlated with the expression of various matrix metalloproteinases (MMPs). Elevated AMC may serve as a predictor of higher risk of TGCTS0, and CCL2 mediated TAMs infiltration and MMPs secretion is essential for the tumorigenesis of TGCT.

Detecting aberrant DNA methylation in Illumina DNA methylation arrays: a toolbox and recommendations for its use

ABSTRACT In this study, our goal was to determine probe-specific thresholds for identifying aberrant, or outlying, DNA methylation and to provide guidance on the relative merits of using continuous or outlier methylation data. To construct a reference database, we downloaded Illumina Human 450K array data for more than 2,000 normal samples, characterized the distribution of DNA methylation and derived probe-specific thresholds for identifying aberrations. We made the decision to restrict our reference database to solid normal tissue and morphologically normal tissue found adjacent to solid tumours, excluding blood which has very distinctive patterns of DNA methylation. Next, we explored the utility of our outlier thresholds in several analyses that are commonly performed on DNA methylation data. Outliers are as effective as the full continuous dataset for simple tasks, like distinguishing tumour tissue from normal, but becomes less useful as the complexity of the problem increases. We developed an R package called OutlierMeth containing our thresholds, as well as functions for applying them to data.

A combined radiomic model distinguishing GISTs from leiomyomas and schwannomas in the stomach based on endoscopic ultrasonography images.

Endoscopic ultrasonography (EUS) is recommended as the best tool for evaluating gastric subepithelial lesions (SELs); nonetheless, it has difficulty distinguishing gastrointestinal stromal tumors (GISTs) from leiomyomas and schwannomas. GISTs have malignant potential, whereas leiomyomas and schwannomas are considered benign. This study aimed to establish a combined radiomic model based on EUS images for distinguishing GISTs from leiomyomas and schwannomas in the stomach. EUS images of pathologically confirmed GISTs, leiomyomas, and schwannomas were collected from five centers. Gastric SELs were divided into training and testing datasets based on random split-sample method (7:3). Radiomic features were extracted from the tumor and muscularis propria regions. Principal component analysis, least absolute shrinkage, and selection operator were used for feature selection. Support vector machine was used to construct radiomic models. Two radiomic models were built: the conventional radiomic model included tumor features alone, whereas the combined radiomic model incorporated features from the tumor and muscularis propria regions. A total of 3933 EUS images from 485 cases were included. For the differential diagnosis of GISTs from leiomyomas and schwannomas, the accuracy, sensitivity, specificity, and area under the receiver operating characteristic curve were 74.5%, 72.2%, 78.7%, and 0.754, respectively, for the EUS experts; 76.8%, 74.4%, 81.0%, and 0.830, respectively, for the conventional radiomic model; and 90.9%, 91.0%, 90.6%, and 0.953, respectively, for the combined radiomic model. For gastric SELs<20mm, the accuracy, sensitivity, specificity, and area under the receiver operating characteristic curve of the combined radiomic model were 91.4%, 91.6%, 91.1%, and 0.960, respectively. We developed and validated a combined radiomic model to distinguish gastric GISTs from leiomyomas and schwannomas. The combined radiomic model showed better diagnostic performance than the conventional radiomic model and could assist EUS experts in non-invasively diagnosing gastric SELs, particularly gastric SELs<20mm.

Deep learning referral suggestion and tumour discrimination using explainable artificial intelligence applied to multiparametric MRI.

An appropriate and fast clinical referral suggestion is important for intra-axial mass-like lesions (IMLLs) in the emergency setting. We aimed to apply an interpretable deep learning (DL) system to multiparametric MRI to obtain clinical referral suggestion for IMLLs, and to validate it in the setting of nontraumatic emergency neuroradiology. A DL system was developed in 747 patients with IMLLs ranging 30 diseases who underwent pre- and post-contrast T1-weighted (T1CE), FLAIR, and diffusion-weighted imaging (DWI). A DL system that segments IMLLs, classifies tumourous conditions, and suggests clinical referral among surgery, systematic work-up, medical treatment, and conservative treatment, was developed. The system was validated in an independent cohort of 130 emergency patients, and performance in referral suggestion and tumour discrimination was compared with that of radiologists using receiver operating characteristics curve, precision-recall curve analysis, and confusion matrices. Multiparametric interpretable visualisation of high-relevance regions from layer-wise relevance propagation overlaid on contrast-enhanced T1WI and DWI was analysed. The DL system provided correct referral suggestions in 94 of 130 patients (72.3%) and performed comparably to radiologists (accuracy 72.6%, McNemar test; p = .942). For distinguishing tumours from non-tumourous conditions, the DL system (AUC, 0.90 and AUPRC, 0.94) performed similarly to human readers (AUC, 0.81~0.92, and AUPRC, 0.88~0.95). Solid portions of tumours showed a high overlap of relevance, but non-tumours did not (Dice coefficient 0.77 vs. 0.33, p < .001), demonstrating the DL's decision. Our DL system could appropriately triage patients using multiparametric MRI and provide interpretability through multiparametric heatmaps, and may thereby aid neuroradiologic diagnoses in emergency settings. Our AI triages patients with raw MRI images to clinical referral pathways in brain intra-axial mass-like lesions. We demonstrate that the decision is based on the relative relevance between contrast-enhanced T1-weighted and diffusion-weighted images, providing explainability across multiparametric MRI data. • A deep learning (DL) system using multiparametric MRI suggested clinical referral to patients with intra-axial mass-like lesions (IMLLs) similar to radiologists (accuracy 72.3% vs. 72.6%). • In the differentiation of tumourousand non-tumourousconditions, the DL system (AUC, 0.90) performed similar with radiologists (AUC, 0.81-0.92). • The DL's decision basis for differentiating tumours from non-tumours can be quantified using multiparametric heatmaps obtained via the layer-wise relevance propagation method.

Abstract 5396: Early detection of ovarian cancer using mRNA sample data to investigate the correlation of specific gene mutations with ovarian cancer

Abstract Introduction: Ovarian cancer (OC) is the second most common gynecologic cancer in the U.S, and is the deadliest cancer of the female reproductive system. Because of the complex nature of OC, diagnoses are often by display of symptoms at late stages of disease. Identifying OC at early stages can increase the 5-year survival rate from 40% in stage IV to 94% in stage I. Therefore, effective screening and detection of OC at early stages is imperative. The main objective of this study is to develop a diagnostic panel of markers that correlate with early-stage OC, and ultimately, improve the survival rate of patients. Background: Approaches in the past several years to detect malignant OC have focused on processing relevant images or screening genomic data. These studies use liquid-based biopsies, sonographic images, cyst images, CT images, MR imaging and genomics data. Some studies have focused on distinguishing malignant tumors versus benign tumor cells. In addition to statistical methods to examine available genomic datasets, Machine Learning models including K-Nearest Neighbor, K-means, Convolutional Neural Networks, Deep Learning, Support Vector Machine, Random Forest, and Fuzzy algorithms have been used. Methods: In this study we proposed a method to determine a possible correlation between some actively expressed genes and OC, using publicly available mRNA data sets. The samples were retrieved from the GSE106817 dataset via the NCBI GEO site, with 333 cancerous and 2759 Non-cancerous samples. The objective is to define a set of genes which is able to illustrate prediction of correlation through the use of PCA, NMF and SVD. We further verified our findings by testing on one additional different mRNA datasets. For this, we compared several machine learning algorithms to potentially correlate active genes and OC stages, from the available data. Results and Discussions: Through our study we have developed Machine Learning models valuable to identify patterns of disease at different stages, thus, potentially be useful in developing predictive models. Our initial experimental results have been demonstrated with a set of mRNA signatures, which may be used to create a panel of biomarkers valuable for detecting early-stage OC. Our results confirm the correlation between OC and PAX8, PEG3, and BIRC5, and MYB and RAC2. Continuation of this study highlights methods towards the early detection of OC and to improve disease identification outcome. Citation Format: Zahra Saghaie, Christine Richardson, M. Taghi Mostafavi. Early detection of ovarian cancer using mRNA sample data to investigate the correlation of specific gene mutations with ovarian cancer. [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 5396.

Evaluation of the Expression Intensity of Glucose Transporter-1 Marker and its Diagnostic Value in Differentiating Between Borderline and Malignant Ovarian Epithelial Tumors

Background: Ovarian cancer ranks second among gynecological cancers worldwide. This study aimed to compare the glucose transporter-1 (GLUT-1) expression in benign, borderline, and malignant ovarian epithelial tumors and evaluate GLUT-1 expression as a diagnostic tool for distinguishing tumors in the ovary. Materials and Methods: This descriptive-analytical cross-sectional study analyzed 69 pathological samples of patients diagnosed with ovarian epithelial tumors who underwent oophorectomy. Immunohistochemical staining was performed using GLUT-1 antibody. The intensity of cell membrane staining and the proportion of positive neoplastic cells were graded to score immunostaining. Chi-square/Fisher’s exact test was used to analyze the data. The sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV), and staining accuracy for GLUT-1 in distinguishing borderline from invasive tumors were calculated by standard methods (P&lt;0.05). Results: In all benign tumors, GLUT-1 staining was negative. In addition, weak staining intensity was observed in 38.5% of borderline tumors, and 96% of invasive tumors had strong staining intensity (P&lt;0.001). Strong GLUT-1 staining was found in 94.7% of Papillary Serous Carcinoma, 9/1% of Borderline Serous tumors, 100% of Brenner tumors, and clear cell carcinoma. The results demonstrated a high diagnostic value of GLUT-1 expression intensity in differentiating between borderline and malignant ovarian epithelial tumors (Accuracy: 97.10, Sensitivity: 96%, Specificity: 97.73). Conclusion: Overall, GLUT-1 expression could help distinguish benign from borderline, especially borderline from malignant ovarian epithelial tumors. Thus, it seems that it provides useful prognostic information, particularly for the borderline category.