scholarly journals Comparison of Comprehensive Morphological and Radiomics Features of Subsolid Pulmonary Nodules to Distinguish Minimally Invasive Adenocarcinomas and Invasive Adenocarcinomas in CT Scan

2022 ◽  
Vol 11 ◽  
Author(s):  
Lu Qiu ◽  
Xiuping Zhang ◽  
Haixia Mao ◽  
Xiangming Fang ◽  
Wei Ding ◽  
...  
Keyword(s):  
Minimally Invasive ◽  
Clinical Decision Making ◽  
High Efficiency ◽  
Pulmonary Nodules ◽  
Roc Curves ◽  
Clinical Decision ◽  
Gray Level ◽  
Combined Model ◽  
Significant Difference ◽  
Morphological Model

ObjectiveTo investigative the diagnostic performance of the morphological model, radiomics model, and combined model in differentiating invasive adenocarcinomas (IACs) from minimally invasive adenocarcinomas (MIAs).MethodsThis study retrospectively involved 307 patients who underwent chest computed tomography (CT) examination and presented as subsolid pulmonary nodules whose pathological findings were MIAs or IACs from January 2010 to May 2018. These patients were randomly assigned to training and validation groups in a ratio of 4:1 for 10 times. Eighteen categories of morphological features of pulmonary nodules including internal and surrounding structure were labeled. The following radiomics features are extracted: first-order features, shape-based features, gray-level co-occurrence matrix (GLCM) features, gray-level size zone matrix (GLSZM) features, gray-level run length matrix (GLRLM) features, and gray-level dependence matrix (GLDM) features. The chi-square test and F1 test selected morphology features, and LASSO selected radiomics features. Logistic regression was used to establish models. Receiver operating characteristic (ROC) curves evaluated the effectiveness, and Delong analysis compared ROC statistic difference among three models.ResultsIn validation cohorts, areas under the curve (AUC) of the morphological model, radiomics model, and combined model of distinguishing MIAs from IACs were 0.88, 0.87, and 0.89; the sensitivity (SE) was 0.68, 0.81, and 0.83; and the specificity (SP) was 0.93, 0.79, and 0.87. There was no statistically significant difference in AUC between three models (p > 0.05).ConclusionThe morphological model, radiomics model, and combined model all have a high efficiency in the differentiation between MIAs and IACs and have potential to provide non-invasive assistant information for clinical decision-making.

scholarly journals Predicting mortality in adults with suspected infection in a Rwandan hospital: an evaluation of the adapted MEWS, qSOFA and UVA scores

BMJ Open ◽  
2021 ◽  
Vol 11 (2) ◽  
pp. e040361
Author(s):  
Amanda Klinger ◽  
Ariel Mueller ◽  
Tori Sutherland ◽  
Christophe Mpirimbanyi ◽  
Elie Nziyomaze ◽  
...  
Keyword(s):  
Decision Making ◽  
Quality Improvement ◽  
Clinical Decision Making ◽  
Clinical Decision ◽  
Tertiary Hospital ◽  
Early Warning Score ◽  
Receiver Operating Curve ◽  
Predictive Capacity ◽  
Suspected Infection ◽  
Significant Difference

RationaleMortality prediction scores are increasingly being evaluated in low and middle income countries (LMICs) for research comparisons, quality improvement and clinical decision-making. The modified early warning score (MEWS), quick Sequential (Sepsis-Related) Organ Failure Assessment (qSOFA), and Universal Vital Assessment (UVA) score use variables that are feasible to obtain, and have demonstrated potential to predict mortality in LMIC cohorts.ObjectiveTo determine the predictive capacity of adapted MEWS, qSOFA and UVA in a Rwandan hospital.Design, setting, participants and outcome measuresWe prospectively collected data on all adult patients admitted to a tertiary hospital in Rwanda with suspected infection over 7 months. We calculated an adapted MEWS, qSOFA and UVA score for each participant. The predictive capacity of each score was assessed including sensitivity, specificity, positive and negative predictive value, OR, area under the receiver operating curve (AUROC) and performance by underlying risk quartile.ResultsWe screened 19 178 patient days, and enrolled 647 unique patients. Median age was 35 years, and in-hospital mortality was 18.1%. The proportion of data missing for each variable ranged from 0% to 11.7%. The sensitivities and specificities of the scores were: adapted MEWS >4, 50.4% and 74.9%, respectively; qSOFA >2, 24.8% and 90.4%, respectively; and UVA >4, 28.2% and 91.1%, respectively. The scores as continuous variables demonstrated the following AUROCs: adapted MEWS 0.69 (95% CI 0.64 to 0.74), qSOFA 0.65 (95% CI 0.60 to 0.70), and UVA 0.71 (95% CI 0.66 to 0.76); there was no statistically significant difference between the discriminative capacities of the scores.ConclusionThree scores demonstrated a modest ability to predict mortality in a prospective study of inpatients with suspected infection at a Rwandan tertiary hospital. Careful consideration must be given to their adequacy before using them in research comparisons, quality improvement or clinical decision-making.

4971Blood viscosity and its relevance to the diagnosis and management of pulmonary hypertension: a new elephant in the cathlab

2019 ◽  
Vol 40 (Supplement_1) ◽  
Author(s):  
A Kempny ◽  
K Dimopoulos ◽  
A E Fraisse ◽  
G P Diller ◽  
L C Price ◽  
...  
Keyword(s):  
Pulmonary Hypertension ◽  
Cardiac Catheterization ◽  
Clinical Decision Making ◽  
Clinical Decision ◽  
Pulmonary Vasculature ◽  
Pulmonary Vascular Disease ◽  
Significant Difference ◽  
The Difference ◽  
Clinically Significant ◽  
The Relationship

Abstract Background Pulmonary vascular resistance (PVR) is an essential parameter assessed during cardiac catheterization. It is used to confirm pulmonary vascular disease, to assess response to targeted pulmonary hypertension (PH) therapy and to determine the possibility of surgery, such as closure of intra-cardiac shunt or transplantation. While PVR is believed to mainly reflect the properties of the pulmonary vasculature, it is also related to blood viscosity (BV). Objectives We aimed to assess the relationship between measured (mPVR) and viscosity-corrected PVR (cPVR) and its impact on clinical decision-making. Methods We assessed consecutive PH patients undergoing cardiac catheterization. BV was assessed using the Hutton method. Results We included 465 patients (56.6% female, median age 63y). The difference between mPVR and cPVR was highest in patients with abnormal Hb levels (anemic patients: 5.6 [3.4–8.0] vs 7.8Wood Units (WU) [5.1–11.9], P<0.001; patients with raised Hb: 10.8 [6.9–15.4] vs. 7.6WU [4.6–10.8], P<0.001, respectively). Overall, 33.3% patients had a clinically significant (>2.0WU) difference between mPVR and cPVR, and this was more pronounced in those with anemia (52.9%) or raised Hb (77.6%). In patients in the upper quartile for this difference, mPVR and cPVR differed by 4.0WU [3.4–5.2]. Adjustment of PVR required Conclusions We report, herewith, a clinically significant difference between mPVR and cPVR in a third of contemporary patients assessed for PH. This difference is most pronounced in patients with anemia, in whom mPVR significantly underestimates PVR, whereas in most patients with raised Hb, mPVR overestimates it. Our data suggest that routine adjustment for BV is necessary.

Standardized Synoptic Cancer Pathology Reports — So What and Who Cares?: A Population-Based Satisfaction Survey of 970 Pathologists, Surgeons, and Oncologists

2013 ◽  
Vol 137 (11) ◽  
pp. 1599-1602 ◽  
Author(s):  
Sara Lankshear ◽  
John Srigley ◽  
Thomas McGowan ◽  
Marta Yurcan ◽  
Carol Sawka
Keyword(s):  
Decision Making ◽  
Clinical Decision Making ◽  
Clinical Decision ◽  
Physician Satisfaction ◽  
Cross Sectional ◽  
Cancer Pathology ◽  
Report Turnaround Time ◽  
Significant Difference ◽  
Pathology Reporting ◽  
The Impact

Context.—Cancer Care Ontario implemented synoptic pathology reporting across Ontario, impacting the practice of pathologists, surgeons, and medical and radiation oncologists. The benefits of standardized synoptic pathology reporting include enhanced completeness and improved consistency in comparison with narrative reports, with reported challenges including increased workload and report turnaround time. Objective.—To determine the impact of synoptic pathology reporting on physician satisfaction specific to practice and process. Design.—A descriptive, cross-sectional design was utilized involving 970 clinicians across 27 hospitals. An 11-item survey was developed to obtain information regarding timeliness, completeness, clarity, and usability. Open-ended questions were also employed to obtain qualitative comments. Results.—A 51% response rate was obtained, with descriptive statistics reporting that physicians perceive synoptic reports as significantly better than narrative reports. Correlation analysis revealed a moderately strong, positive relationship between respondents' perceptions of overall satisfaction with the level of information provided and perceptions of completeness for clinical decision making (r = 0.750, P &lt; .001) and ease of finding information for clinical decision making (r = 0.663, P &lt; .001). Dependent t tests showed a statistically significant difference in the satisfaction scores of pathologists and oncologists (t169 = 3.044, P = .003). Qualitative comments revealed technology-related issues as the most frequently cited factor impacting timeliness of report completion. Conclusion.—This study provides evidence of strong physician satisfaction with synoptic cancer pathology reporting as a clinical decision support tool in the diagnosis, prognosis, and treatment of cancer patients.

scholarly journals Use of the Tissue Common Rejection Module Score in Kidney Transplant as an Objective Measure of Allograft Inflammation

2021 ◽  
Vol 11 ◽  
Author(s):  
Arya Zarinsefat ◽  
Jose M. Arreola Guerra ◽  
Tara Sigdel ◽  
Izabella Damm ◽  
Reuben Sarwal ◽  
...  
Keyword(s):  
Kidney Transplant ◽  
Clinical Decision Making ◽  
Clinical Sample ◽  
Rna Extraction ◽  
Prospective Trial ◽  
Objective Measure ◽  
Clinical Decision ◽  
Biopsy Specimens ◽  
Significant Difference ◽  
Quantitative Score

Long-term kidney transplant (KT) allograft outcomes have not improved as expected despite a better understanding of rejection and improved immunosuppression. Previous work had validated a computed rejection score, the tissue common rejection module (tCRM), measured by amplification-based assessment of 11 genes from formalin-fixed paraffin-embedded (FFPE) biopsy specimens, which allows for quantitative, unbiased assessment of immune injury. We applied tCRM in a prospective trial of 124 KT recipients, and contrasted assessment by tCRM and histology reads from 2 independent pathologists on protocol and cause biopsies post-transplant. Four 10-μm shaves from FFPE biopsy specimens were used for RNA extraction and amplification by qPCR of the 11 tCRM genes, from which the tCRM score was calculated. Biopsy diagnoses of either acute rejection (AR) or borderline rejection (BL) were considered to have inflammation present, while stable biopsies had no inflammation. Of the 77 biopsies that were read by both pathologists, a total of 40 mismatches in the diagnosis were present. The median tCRM scores for AR, BL, and stable diagnoses were 4.87, 1.85, and 1.27, respectively, with an overall significant difference among all histologic groups (Kruskal-Wallis, p &lt; 0.0001). There were significant differences in tCRM scores between pathologists both finding inflammation vs. disagreement (p = 0.003), and both finding inflammation vs. both finding no inflammation (p &lt; 0.001), along with overall significance between all scores (Kruskal-Wallis, p &lt; 0.001). A logistic regression model predicting graft inflammation using various clinical predictor variables and tCRM revealed the tCRM score as the only significant predictor of graft inflammation (OR: 1.90, 95% CI: 1.40–2.68, p &lt; 0.0001). Accurate, quantitative, and unbiased assessment of rejection of the clinical sample is critical. Given the discrepant diagnoses between pathologists on the same samples, individuals could utilize the tCRM score as a tiebreaker in unclear situations. We propose that the tCRM quantitative score can provide unbiased quantification of graft inflammation, and its rapid evaluation by PCR on the FFPE shave can become a critical adjunct to help drive clinical decision making and immunosuppression delivery.

scholarly journals Differences in spatial distribution between WHO 2016 low-grade glioma molecular subgroups

2019 ◽  
Vol 1 (1) ◽  
Author(s):  
Maarten M J Wijnenga ◽  
Sebastian R van der Voort ◽  
Pim J French ◽  
Stefan Klein ◽  
Hendrikus J Dubbink ◽  
...  
Keyword(s):  
Decision Making ◽  
Spatial Distribution ◽  
Clinical Decision Making ◽  
Tumor Volume ◽  
Right Hemisphere ◽  
Clinical Decision ◽  
Low Grade Glioma ◽  
Low Grade ◽  
Wild Type ◽  
Significant Difference

Abstract Background Several studies reported a correlation between anatomic location and genetic background of low-grade gliomas (LGGs). As such, tumor location may contribute to presurgical clinical decision-making. Our purpose was to visualize and compare the spatial distribution of different WHO 2016 gliomas, frequently aberrated single genes and DNA copy number alterations within subgroups, and groups of postoperative tumor volume. Methods Adult grade II glioma patients (WHO 2016 classified) diagnosed between 2003 and 2016 were included. Tumor volume and location were assessed with semi-automatic software. All volumes of interest were mapped to a standard reference brain. Location heatmaps were created for each WHO 2016 glioma subgroup, frequently aberrated single genes and copy numbers (CNVs), as well as heatmaps according to groups of postoperative tumor volume. Differences between subgroups were determined using voxelwise permutation testing. Results A total of 110 IDH mutated astrocytoma patients, 92 IDH mutated and 1p19q co-deleted oligodendroglioma patients, and 22 IDH wild-type astrocytoma patients were included. We identified small regions in which specific molecular subtypes occurred more frequently. IDH-mutated LGGs were more frequently located in the frontal lobes and IDH wild-type tumors more frequently in the basal ganglia of the right hemisphere. We found no localizations of significant difference for single genes/CNVs in subgroups, except for loss of 9p in oligodendrogliomas with a predilection for the left parietal lobes. More extensive resections in LGG were associated with frontal locations. Conclusions WHO low-grade glioma subgroups show differences in spatial distribution. Our data may contribute to presurgical clinical decision-making in LGG patients.

scholarly journals Influence of breast reconstruction on technical aspects of echocardiographic image acquisition compared with physician-assessed image quality

2019 ◽  
Vol 5 (1) ◽  
Author(s):  
Joaquin Duarte Ow ◽  
Mohamad Hemu ◽  
Anel Yakupovich ◽  
Parva Bhatt ◽  
Hannah Gaddam ◽  
...  
Keyword(s):  
Breast Cancer ◽  
Decision Making ◽  
Image Quality ◽  
Cardiac Function ◽  
Breast Reconstruction ◽  
Clinical Decision Making ◽  
Image Acquisition ◽  
Clinical Decision ◽  
Echocardiographic Image ◽  
Significant Difference

Abstract Introduction Assessment of cardiac function after treatment for breast cancer relies on interval evaluation of ventricular function through echocardiography. Women who undergo mastectomy more frequently choose to undergo breast reconstruction with implant. This could impede assessment of cardiac function in those with left-sided implant. We aimed to examine whether left-sided breast reconstruction with tissue expanders (TE) affect echo image acquisition and quality, possibly affecting clinical decision-making. Methods A retrospective case-control study was conducted in 190 female breast cancer patients who had undergone breast reconstruction with TE at an urban academic center. Echocardiographic technical assessment and image quality were respectively classified as excellent/good or adequate/technically difficult by technicians; and excellent/good or adequate/poor by 2 board-certified cardiologist readers. Likelihood ratio was used to test multivariate associations between image quality and left-sided TE. Results We identified 32 women (81.3% white; mean age 48 years) with left-sided/bilateral TE, and 158 right-sided/no TE (76.6% white, mean age 57 years). In multivariable analyses, we found a statistically significant difference in technician-assessed difficulty in image acquisition between cases and controls (p = 0.01); but no differences in physician-assessed image quality between cases and controls (p = 0.09, Pearson’s r = 0.467). Conclusions Left-sided breast TE appears to affect the technical difficulty of echo image acquisition, but not physician-assessed echo image quality. This likely means that echo technicians absorb most of the impediments associated with imaging patients with breast TE such that the presence of TE has no bearing on downstream clinical decision-making associated with echo image quality.

Patterns of palliative radiation near the end of life: A single-institution retrospective analysis.

2013 ◽  
Vol 31 (15_suppl) ◽  
pp. 9636-9636 ◽  
Author(s):  
Sara R. Alcorn ◽  
Thomas J. Smith ◽  
Todd R. McNutt ◽  
M. Jennifer Cheng ◽  
Sydney Morss Dy ◽  
...  
Keyword(s):  
End Of Life ◽  
Retrospective Analysis ◽  
Clinical Decision Making ◽  
Clinical Decision ◽  
Johns Hopkins Hospital ◽  
National Guidelines ◽  
Single Fraction ◽  
Palliative Intent ◽  
Risk Of Death ◽  
Significant Difference

9636 Background: The care of patients who receive radiation therapy (RT) at the end of life (EOL) is under scrutiny to ensure effectiveness and value, with many patients not completing RT (Gripp, 2010; Toole, 2012). This retrospective analysis seeks to describe patterns of utilization of palliative RT, including rates of completion of RT offered at the EOL and the use of single fraction RT for bone metastases. Methods: Electronic medical records were used to create a database of 3,383 RT plans for brain, bone, lung, and other metastatic sites in patients treated at Johns Hopkins Hospital from 9/1/2007-7/15/2012. RT plans without palliative intent were excluded. T-tests and logistic regression compared patient and treatment characteristics between patients who died > 1 month versus ≤ 1 month after their last RT fraction. Results: A total of 983 patients were treated to 1,524 sites, with an average of 1.7 RT sites (SD 1.3) per patient. Of these, 872 (89%) patients had complete records and were included in analysis. At the time of analysis, 85% had died. The mean age of 62.1 years (SD 3.4) did not differ statistically based on time from RT to death. Death ≤ 1 month after RT was documented in 215 (24.7%) patients. Compared to patients living > 1 month after RT, patients receiving RT within the last month of life were more likely to be lung (17% versus 9%), less likely to be brain (34% versus 44%), and equally likely to be bone (45% versus 43%) sites. Patients who died ≤ 1 month after completing RT spent on average 5 days (16.6%) of the last month of life receiving RT, with no significant difference by disease site. Conclusions: Most patients receiving palliative RT finish therapy, with 25% dying ≤ 1 month after RT. Single fraction bone RT was relatively uncommon, with no significant difference in the rates of single fraction RT based on time from RT to death. These data provide a framework to match treatment patterns with national guidelines. Additionally, they provide context to model risk of death shortly after RT, which can aid in clinical decision-making. [Table: see text]

scholarly journals 1629 The Risk of Malignancy in Cytologically Indeterminate Thyroid Nodules

2021 ◽  
Vol 108 (Supplement_6) ◽  
Author(s):  
O Jones ◽  
V Blackabey ◽  
N Bhat
Keyword(s):  
Thyroid Nodules ◽  
Clinical Decision Making ◽  
Clinical Decision ◽  
Needle Aspiration ◽  
Diagnostic Assessment ◽  
Patient Counselling ◽  
Conversion Rates ◽  
Significant Difference ◽  
Risk Of Malignancy ◽  
Malignant Conversion

Abstract Introduction Fine needle aspiration (FNA) is the gold standard for the diagnostic assessment of thyroid nodules, with cytology stratified using the Thy classification (Thy 1-5). The management of cases where cytology is indeterminate (Thy-3) is challenging and subject to controversy. The current British Thyroid Association guidelines subclassify Thy-3 lesions into Thy-3a (atypia) and Thy-3f (follicular). Repeat FNA is generally recommended for Thy-3a specimens, whilst Thy-3f lesions should proceed to diagnostic hemithyroidectomy. The aim of this study was to determine the risk of malignancy in Thy-3a and Thy-3f lesions. Method This was a retrospective study of all patients who underwent FNA of a thyroid nodule from 01/01/2018 – 31/12/19. Those with Thy-3 cytology were identified and results correlated with final surgical histology. Results In total, there were 179 patients with Thy-3 cytology: 37 Thy-3a and 142 Thy-3f. The rate of malignancy was 21.6% (n = 8) for Thy-3a lesions and 20.4% (n = 29) for Thy-3f. When excluding microcarcinoma, the rates fell to 20.4% (n = 7) and 12.7% (n = 18) respectively. There was no statistically significant difference in malignant conversion between the two groups (p = 0.20). Conclusions This study does not demonstrate any statistically significant difference in the risk of malignancy between the Thy-3a and Thy-3f groups, suggesting that this subclassification does not add any additional utility to clinical decision making. Consideration should be given to the use of local malignant conversion rates to guide further management and pre-operative patient counselling in the Thy3 group.

scholarly journals 1157. Determining the Clinical Utility of 16S rRNA in the Management of Pediatric Infections

2021 ◽  
Vol 8 (Supplement_1) ◽  
pp. S670-S670
Author(s):  
Peter Paul Lim ◽  
Ankita P Desai ◽  
Sree Sarah Cherian ◽  
Sindhoosha Malay
Keyword(s):  
Decision Making ◽  
16S Rrna ◽  
Clinical Utility ◽  
Clinical Decision Making ◽  
Clinical Decision ◽  
Joint Fluid ◽  
Significant Difference ◽  
Pediatric Infections ◽  
Tissue Specimens ◽  
Antimicrobial Regimen

Abstract Background Conventional culture remains the gold standard to facilitate a targeted antimicrobial regimen in the treatment of bacterial infections. However, certain pediatric infections are caused by fastidious organisms and treatment with antibiotics prior to specimen collection may hamper growth of pathogens in routine culture. The use of 16S rRNA in culture negative infections has improved identification of bacterial pathogens in select scenarios. However, the specific impact of 16S rRNA on clinical decision making, especially in pediatric infections, is not well-defined. This study aims to elucidate the utility of 16S rRNA on clinical management of pediatric infections. Methods A retrospective analysis was done on different clinical specimens which had 16S rRNA performed from August 2016 – March 2020 in our institution. Detailed chart review was performed to determine how the 16S rRNA result impacted clinical decision making. Clinical utility was defined as change in patient’s overall antimicrobial regimen, pathogen confirmation, and treatment duration. Results Seventy-four samples from 71 pediatric patients were included in the analysis: 32 (43%) were fluid specimens and 42 (57%) were tissue specimens. Significant clinical utility was identified in 30 (40.5%) of 74 clinical samples (p &lt; 0.0001). Of all specimens, pulmonary samples yielded the most clinical utility (n=9, 30%) followed equally by joint fluid (n=6, 20%) and bone (n=6, 20%). There was no significant difference in clinical utility between fluid and tissue specimens (p= 0.346). In 64 patients whose antimicrobial spectrum coverage was analyzed, patients with broad spectrum coverage was decreased from 48 to 21 and narrow spectrum coverage increased from 16 to 43 using 16S rRNA result, though not significant (p= 0.4111). Of all patients included in the analysis, the median number of antibiotics used before 16S rRNA result, 2, was significantly decreased to 1 (p &lt; 0.0001). Conclusion 16S rRNA has a significant impact in terms of decreasing number of antibiotics used in treatment of pediatric infections. Pulmonary specimens have the highest clinical utility among all samples. Additional cost benefit analysis needs to be completed to further determine clinical benefit. Disclosures All Authors: No reported disclosures

scholarly journals CT Radiomics and Morphologic Characteristics for Predicting PD-L1 Expression on Tumour Cells and Tumour Infiltrating Lymphocytes in Gastric Cancer

2020 ◽  
Author(s):  
Xiangmei Qiao ◽  
Lin Li ◽  
Zhengliang Li ◽  
Changfeng Ji ◽  
Hui Li ◽  
...  
Keyword(s):  
Gastric Cancer ◽  
Clinical Decision Making ◽  
Roc Curves ◽  
Clinical Decision ◽  
Tumour Cells ◽  
Support Vector ◽  
Diagnostic Efficiency ◽  
Morphologic Characteristics ◽  
Tumour Infiltrating Lymphocytes ◽  
Infiltrating Lymphocytes

Abstract Background To explore CT radiomics and morphologic characteristics for predicting programmed cell death ligand 1 on tumour cells (PD-L1) and tumour infiltrating lymphocytes (PD-L1-TILs) status in gastric cancer (GC). Methods From March 2019 to October 2019, 101 patients identified with GCs who underwent surgery at our hospital were enrolled in this study retrospectively. Radiomic features were extracted from regions of interest manually drawn on venous CT images. Besides, 13 morphologic characteristics were evaluated. The signatures based on radiomics and morphologic characteristics were built using multiple classifiers (Support Vector Machine [SVM], Naive Bayes [NB], Decision Trees [DT], and Random Forest [RF]). Receiver operating characteristic (ROC) curve was performed to assess diagnostic efficiency. Results The adjacent adipose tissue (P = 0.009) and numerous radiomic features (all P < 0.05) differed significantly between GCs with different PD-L1 status. Six radiomic features showed significant differences between different PD-L1-TILs status (all P < 0.05). The highest areas under the ROC curves (AUCs) of signatures generated by classifiers were 0.807 (SVM) and 0.729 (NB) for the prediction of PD-L1 and PD-L1-TILs status, respectively. Conclusions It was promising to predict PD-L1 status in GCs noninvasively using CT radiomics combined with morphologic characteristics. It might help to improve clinical decision making with regard to immunotherapy. However, the prediction for PD-L1-TILs needs to be explored further.

Sign in / Sign up

Export Citation Format

Share Document