Research Presentation Session: Chest

RPS 1304 - Advanced imaging of thoracic oncologic patients

March 1, 09:30 - 11:00 CET

7 min
DEB-BACE followed by systemic chemotherapy vs systemic chemotherapy alone for advanced lung adenocarcinoma: a propensity score match study
Linqiang Lai, Lishui / China
Author Block: J. Tu; Lishui/CN
Purpose: This retrospective study aimed to investigate the effectiveness and safety of sequential bronchial arterial chemoembolisation with drug-eluting beads (DEB-BACE) and chemotherapy versus systemic chemotherapy alone for advanced lung adenocarcinoma progressing on targeted therapy.
Methods or Background: Stage III or IV lung adenocarcinoma patients in the chemotherapy group received an intravenous injection of pemetrexed 500 mg/m2 plus cisplatin 75 mg/m2 on day 1 for 4 cycles, with each cycle lasting for 21 days. Patients in the DEB-BACE plus chemotherapy group underwent DEB-BACE using CalliSpheres drug-eluting beads loaded with gemcitabine (400 mg) and received via the microcatheter cisplatin 75 mg/m2 and gemcitabine 600 mg/m2 followed by intravenous chemotherapy 3 weeks post DEB-BACE. The primary outcome was overall survival.
Results or Findings: Sixty-two patients who received chemotherapy and 69 who received DEB-BACE plus chemotherapy were included, with 36 patients in each group after PSM. After PSM, the median OS was 18.3 months in the chemotherapy group and 33.1 months in the DEB-BACE plus chemotherapy group (P < 0.001). Multivariate Cox regression analysis showed that DEB-BACE plus chemotherapy was associated with an 82% reduction in the risk of death versus chemotherapy only (P < 0.001). After PSM, treatment-emergent adverse events of grade 3 or worse occurred in 2 of 36 patients in the DEB-BACE plus chemotherapy group and 16 of 36 patients in the chemotherapy alone group.
Conclusion: DEB-BACE plus chemotherapy improves the response rate and extends the survival of III-IV lung adenocarcinoma patients progressing on targeted therapy. This offers this patient population, who otherwise have a rather dismal clinical outcome, an effective and safe treatment option.
Limitations: DEB-BACE has a learning curve in terms of target tumour feeding vessel selection, the number of embolised vessels, levels of embolisation and size of embolisation beads.
Funding for this study: Funding was received from the Medical and Health Science and Technology Plan of Zhejiang Province (grant number WKJ-ZJ-1932).
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: The study was approved by the Research Ethics Group of the Ethics Committee at Lishui Central Hospital: ethical review of research (2022) number 373.
7 min
Tree-based models for predicting clinically significant pneumothorax in patients undergoing percutaneous coaxial core lung biopsy: a retrospective cohort study
Miguel Emilio Chevasco Hanze, Barcelona / Spain
Author Block: M. E. Chevasco Hanze, D. Castellon Plaza, S. A. Bolivar, B. Del Rio Carrero, H. H. J. Jofre; Barcelona/ES
Purpose: This study aimed to create a prediction model for the development of significant pneumothorax following a CT-guided coaxial core lung biopsy (CT-CCLB) by employing machine learning tree-based models.
Methods or Background: A total of 469 patients who underwent CT-CCLB were retrospectively included. A list of 22 patient, procedure and lesion characteristics were retrieved. Boruta analysis was used for selection of feature predictors. Afterwards, four tree models, namely CART, AdaBoost, GB and XGBoost, were applied. The final model was chosen based on PPR, PLR and AUC values. Final model and predictors behaviour were further evaluated by tree plot and SHAP analysis.
Results or Findings: Significant pneumothorax rate was 12.79%. GB classifier was found to have the best discriminating power (AUC = 76.82%; PPR = 3.80; PLR = 5.08 43.48%). The top five predictors were lesion size/depth, DLCO, BMI and involvement of fissures/bullae/emphysema during biopsy. Scenarios for encountering the highest significant pneumothorax occur when: 1) the mentioned surfaces are compromised and the procedure is done with the patient in a supine or lateral body position; 2) lesion size is lower than 22.37 mm and BMI is lower than 26.5; 3) lesion size is higher than 22.37 mm and depth is higher than 36.86 mm.
Conclusion: Significant pneumothorax after CT-CCLB was more likely to develop among patients with small lesions, lower BMI, higher depth, and biopsies done in prone/lateral position and with fissures/bullae/emphysema. Machine-learning models demonstrated a high predictive performance, with results being easy to visualise and read.
Limitations: Outweighing outcome was variable. Biopsies were made by radiology residents and attending radiologists
Funding for this study: No funding was received for this study.
Has your study been approved by an ethics committee? Not applicable
Ethics committee - additional information: No ethics committee approval was required because the present study was retrospective, with no intervention throughout the research.
7 min
CT-guided core needle biopsy is safe and accurate for the assessment of pulmonary lesions associated with cystic airspaces
Maurizio Balbi, Turin / Italy
Author Block: M. Balbi1, N. C. Culasso1, M. Barba1, R. Senkeev1, S. Capelli2, A. Caroli2, F. Filipello3, L. Righi1, A. Veltri1; 1Orbassano/IT, 2Bergamo/IT, 3Verduno/IT
Purpose: This study aimed to evaluate the safety and diagnostic capability of CT-guided core needle biopsy (CNB) in pulmonary lesions associated with cystic airspaces (PLACAs).
Methods or Background: Consecutive pulmonary biopsies performed at the San Luigi Gonzaga Hospital (Orbassano, Italy) from February 2010 to January 2022 (n=3069) were retrospectively reviewed to identify patients who underwent CNB for PLACAs (n=90, case group; median age, 69.5 years, 95% confidence interval [CI], 62.0-75.0; 28 females). A group of CNB patients with non-cystic lesions matched for age, sex, emphysema, and lesion depth and dimensions (n=180, control group) was selected to compare the diagnostic yield and complication rate. The diagnostic performance for the final diagnosis was calculated. Univariate and multivariate logistic regressions were performed in case patients to identify risk factors for complications and a non-diagnostic specimen (i.e., nonspecific benignity, atypical cells, insufficient specimen). PLACAs’ specimens were reviewed to assess histopathology.
Results or Findings: There were no significant differences between cases and controls in complication rate (overall: 40% versus 38%; major: 4% versus 6%, respectively) and non-diagnostic specimens (12% versus 9%). The diagnostic performance was similar in both groups (accuracy: 97.78% vs. 97.78%, sensitivity: 97.53% vs. 97.63%, specificity: 100% vs. 100%). Among the patient, procedural, and lesion-related data, the length of the needle pathway through the lung (odds ratio, [OR], 2.86; [95% CI, 1.08-7.80]; p=0.036) and the procedure time (OR, 10.93; 95% CI, 3.77-35.85; p < 0.001) were significant risk factors for complications. No variables predicted a non-diagnostic specimen. In most cases, PLACAs were adenocarcinoma (54%), and the cystic airspaces corresponded to tumour cystification (22 out of 31 resected specimens, 71%).
Conclusion: CT-guided CNB was safe and effective for assessing PLACAs. A long needle pathway and procedure time increased the complication risk.
Limitations: Identified limitations were (1) that this was a single-centre study and (2) the limited number of cases.
Funding for this study: The authors state that this work has not received any funding.
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: The ethics committee approved the present study and waived the need for written informed consent.
7 min
Haemorrhage risk prediction after computed tomography-guided lung biopsy: combining clinical parameters and quantitative pulmonary vascular analysis
Keng-Chian Lin, Taipei / Taiwan, Chinese Taipei
Author Block: K-C. Lin, Y-S. Huang, Y-C. Chang; Taipei/TW
Purpose: This study aimed to evaluate the utility of combining quantitative pulmonary vasculature measures with clinical factors for predicting pulmonary haemorrhage after computed tomography (CT)-guided lung biopsy.
Methods or Background: Patients who underwent CT-guided lung biopsy were retrospectively included in this study. Clinical and radiographic variables were evaluated as predictors of pulmonary haemorrhage. The radiographic pulmonary vascular analysis included vessel count, vessel density, vessel diameter, vessel area, blood volume in small vessels with a cross-sectional area ≤ 5 mm2 (BV5), and total blood vessel volume (TBV) in the lungs. Univariate and multivariate logistic regressions were used to identify the independent risk factors of higher-grade pulmonary haemorrhage and establish the prediction model, which was presented in the form of a nomogram.
Results or Findings: A total of 126 patients was included (discovery cohort n=103, validation cohort n=23). Any pulmonary haemorrhage, higher-grade (grade ≥2) pulmonary haemorrhage, and hemoptysis occurred in 42.9%, 15.9%, and 3.2% of patients who underwent CT-guided lung biopsies. In the discovery cohort, patients with greater lesion depth (p=0.013), higher vessel density (p=0.033), and higher BV5 (p=0.039) were more likely to experience higher-grade haemorrhage. The nomogram prediction model for higher-grade haemorrhage built by the discovery cohort showed similar performance in the validation cohort.
Conclusion: Higher-grade pulmonary haemorrhage may occur after CT-guided lung biopsy. Lesion depth, vessel density, and BV5 are independent risk factors for higher-grade pulmonary haemorrhage. Nomograms that integrate both clinical parameters and radiographic pulmonary vasculature measures offer enhanced capability for the assessment of haemorrhage risk following CT-guided lung biopsy, thereby facilitating improved clinical care for patients.
Limitations: No limitations were identified.
Funding for this study: No funding was received for this study.
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: This study was approved by the Institutional Research Ethics Committee: approval number 202306051RIN.
7 min
De novo low-dose CT-guided lung biopsy technique: minimising radiation with maintained safety and diagnostic yield rates
Avik Banerjee, Leicester / United Kingdom
Author Block: S. Vijayakumar1, A. Banerjee2, G. Tsaknis2; 1Leicester/UK, 2Kettering/UK
Purpose: The purpose of this study is to present our innovative approach in developing a de novo low-dose CT-guided lung biopsy technique, aiming to significantly reduce radiation exposure (DLP less than 100 mGy-cm, only slightly more than CT fluoroscopy) without compromising histopathological accuracy or increasing complication rates.
Methods or Background: We conducted a retrospective analysis of lung biopsy procedures performed using our novel low-dose CT-guided biopsy technique. A total of 100 patients with suspected lung lesions were included. The procedure involved precise planning and real-time image guidance, striking a balance between kVp and mAs to ensure that diagnostic image quality is maintained without unnecessary radiation exposure to the patient. Patient demographics, lesion characteristics, procedural details, radiation dose, histopathological findings, and post-procedural complications were meticulously recorded and analysed.
Results or Findings: The detailed findings are still in process, but some of the highlights include: 1) over 90-95% of cases had DLP below 100 mGy·cm, with almost all cases under 150 mGy·cm; and 2) varied kVp and mAs permutations were explored. Our protocol utilised 100kVp and 50mA, dropped to 30mA during biopsy, utilising precise planning and positions for challenging lung lesions, effectively reducing radiation doses.
Conclusion: Our study introduces a novel low-dose CT-guided lung biopsy technique, reducing radiation exposure to DLP <100mGy·cm while preserving high histopathological accuracy and patient safety. This approach is especially significant for patients needing recurrent scans and biopsies. Its effectiveness positions it as a potential standard, enhancing healthcare quality.
Limitations: For patients with very high BMI, increased radiation dose was necessary to target the lesion, usually higher than our target of 100kVp.
Funding for this study: No funding was received for this study.
Has your study been approved by an ethics committee? Not applicable
Ethics committee - additional information: No information provided by the submitter.
7 min
Adherence to CT surveillance guidelines in early stage post-treatment lung cancer recurrence
Sian Kneafsey, Dublin / Ireland
Author Block: S. Kneafsey, D. P. Garrahy, D. Byrne, P. Beddy; Dublin/IE
Purpose: Detection of lung cancer recurrence in patients post treatment can result in survival benefit, and enable early detection of new primary lung cancer. However, adherence to post-treatment surveillance protocols is highly variable. We assessed the adherence to CT surveillance guidelines in all patients treated with curative resection for non-small cell lung cancer (NSCLC) at our tertiary referral lung cancer centre.
Methods or Background: All patients treated with curative intent surgery are followed up in a nurse led outpatient clinic. Patient demographics, postoperative imaging studies, histopathological results and follow-up appointment schedule were recorded. We defined the minimal standard of surveillance imaging studies (MSSIS) as ≥5 CT studies in the first 5 years as specified in our follow-up guidelines, based on the ESMO guidelines.
Results or Findings: A total of 1243 patients were included. The mean age was 66 ± 9.7. Of these, 579 (46.6%) were male, 665 (53.4%) were female. In total, 787 (63.3%) achieved the minimal standard of surveillance imaging studies directly within our centre, and are undergoing ongoing active surveillance. A further 170 (13.6%) patients achieved the minimal standard of surveillance imaging studies and were discharged from the surveillance service at five years post-surgery. A further 127 (10.2%) patients were discharged to surveillance services in satellite referral hospitals. A total of 149 (12%) patients achieved MSSIS but have since died. Five (0.4%) patients achieved MSSIS but are no longer under surveillance due to palliation. Five (0.4%) patients are no longer under active surveillance due to unknown reasons.
Conclusion: A total of 1111 (89.3%) patients achieved MSSIS at our institution during this surveillance period. Over two-thirds (63.6%) of patients are undergoing active surveillance. Existence of a dedicated nurse-lead lung cancer surveillance outpatient clinic had led to higher surveillance adherence rates than international literature reports.
Limitations: No limitations were identified.
Funding for this study: No funding was received for this study.
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: This study was approved by St. James Hospital Research and Innovation Office: project 8410.
7 min
Coronary calcification and interstitial lung disease are both independently associated with increased mortality in patients undergoing radiotherapy for stage 3 non-small cell lung cancer
Emily Hughes, Glasgow / United Kingdom
Author Block: C. P. McKeag1, E. Hughes1, E. McGarry1, S. Ghatorae2, G. Cowell1, J. Maclay1; 1Glasgow/UK, 2Larbert/UK
Purpose: Patients diagnosed with lung cancer are often multi-morbid. Assessment of comorbidities is possible on diagnostic CT and may influence survival.
Methods or Background: We looked at overall survival outcomes in patients diagnosed with stage 3 non-small cell lung cancer undergoing radical radiotherapy (+/- systemic treatment) in the West of Scotland between 2017-2020. Mortality was right censored at two years, and the cohort consisted of 431 individuals. This was an update to a previous review of outcomes 2017-2019.
We reviewed the diagnostic CT to identify common comorbidities including emphysema, coronary artery calcification, interstitial lung disease, and pleural effusion. Each of these was individually assessed for any impact on survival, and Kaplan-Meier curves were generated, with the aim of identifying potential markers of increased mortality.
Results or Findings: We showed a significant increase in overall mortality for patients with interstitial lung disease (ILD) (p<0.005), and for patients with severe coronary artery calcification compared to those with mild or less calcification (p<0.005). There was no statistically significant difference in mortality between moderate and mild or less coronary calcification (p=0.13), nor between moderate and severe (p=0.23).
There was no statistically significant difference in mortality in patients with or without pleural effusion (p=0.37), nor between severe emphysema and mild or no emphysema (p=0.49).
Conclusion: Severe coronary artery disease and ILD are associated with reduced overall survival in patients undergoing radical radiotherapy for NSCLC. Further studies investigating cardiovascular complications of radiotherapy are required.
Limitations: No limitations were identified.
Funding for this study: No funding was received for this study
Has your study been approved by an ethics committee? Not applicable
Ethics committee - additional information: This was a retrospective study.
7 min
Radiogenomics relationship of non-small cell lung cancer: preliminary results
Maria Paola Belfiore, Naples / Italy
Author Block: M. P. Belfiore, M. Sansone, V. Patanè, R. Monti, F. Grassi, G. Ciani, R. Grassi, S. Cappabianca; Naples/IT
Purpose: The aim of this study was to correlate the radiomics features with the genetic results obtained from liquid biopsy in patients with lung tumours.
Methods or Background: We included 53 patients suffering from NSCLC who underwent pre-surgery CT (GE Revolution 128 MDCT) at the Radiology Department of the Campania University. Every patient performed liquid biopsy subject to informed consent for the genetic analysis. For the radiomic analysis, image processing CT volumes were manually delineated using ITK-snap 3.8.0. Radiomics features (first order, GLCM, GLRLM, GLSZM, GLDM, NGTDM) were computed using Pyradiomics in Python 3.7 environment. For the statistical analysis, association between radiomic features and gene mutations were assessed using feature importance based on ROC analysis; moreover, a machine learning approach based on SVM was used to evaluate the ability of radiomic features to predict gene mutations.
Computations have been performed in the R environment using the CARET package.
Results or Findings: From the genetic analysis it turns out that the accuracy, i.e. the number of correct predictions and the total number of patients, obtainable using the selected group of features is of the order of 0.67. Some correlations between gene and features were found to be the case: ROS.miss.6.43.Arg167Gln with a feature group that included first-order glcm and glszm, ROS-miss.42.43.Asp2213 Asn with a broader set of features and ALK.miss.29.29.Asp1529 Glu with the same previous group but less intense.
Conclusion: Radiomics could better determine the accuracy of malignancy of pulmonary nodules, which have been detected by CT scan in order to treat curatively, select patients with early-stage lung cancer who are appropriate for post-surgical treatment, and determine patients with stage III NSCLC who can tolerate immunotherapy as consolidation therapy after concurrent treatment with chemotherapy-radiation therapy.
Limitations: These results are preliminary and require a greater number of observations.
Funding for this study: No funding was received for this study.
Has your study been approved by an ethics committee? Not applicable
Ethics committee - additional information: No information provided by the submitter.
7 min
A decade of image-guided pleural biopsies: a multicentre retrospective study examining diagnostic yield and complications
Liam Peng, Glasgow / United Kingdom
Author Block: L. Peng, S. Tsim, K. Blyth, G. Cowell; Glasgow/UK
Purpose: Pleural disease is common, with diagnostic options including pleural effusion aspiration, local anaesthetic thoracoscopy (LAT) and image guided biopsy. Despite diagnostic and therapeutic applications, LAT availability varies and is unfeasible with a complex pleural space. Ultrasound (US-) and Computed Tomography (CT-) guided biopsy remain commonly performed, yet post-procedural data is relatively sparse. A high prevalence of malignant pleural mesothelioma (MPM) in the West of Scotland offers an excellent opportunity to assess complication rate and diagnostic yield associated with image guided biopsy relative to other investigation strategies, to shape the consent process, develop care models and aid decision making around timely MPM diagnosis.
Methods or Background: US- and CT-guided pleural biopsy procedures performed in six hospitals between 01.01.13 and 31.03.23 were identified by searching study codes and a key word search. Patient demographics, biopsy pathology, final diagnosis (either by repeat biopsy or consensus diagnosis via regional/national multidisciplinary team meeting) and procedural complications were recorded.
Results or Findings: 194 CT-guided pleural biopsies were performed using a coaxial technique. Seven (3.6%) had a peri-procedural pneumothorax on CT, with two (1.0%) evident on post-procedural chest radiograph. None required intervention. Two cases (1.0%) had haemoptysis. For diagnosis of pleural malignancy, sensitivity of CT-guided biopsy was 93.9% (95% CI 88.9-97.0%), specificity 100% (89.1-100%), positive predictive value (PPV) 100% (97.6-100%) and negative predictive value (NPV) 76.2 (63.7-85.4%).
A total of 79 US-guided pleural biopsies were performed, without post-procedural pneumothorax. For diagnosis of pleural malignancy, sensitivity of US-guided biopsy was 93.1% (95% CI 84.5-97.7%), specificity 100% (59.0-100%), PPV 100% (94.6-100%) and NPV 58.3% (37.5-76.5%).
Conclusion: CT- and US-guided pleural biopsy is safe, offering excellent sensitivity and PPV in the diagnosis of pleural malignancy if LAT is not available or feasible.
Limitations: No limitations were identified.
Funding for this study: No funding was received for this study.
Has your study been approved by an ethics committee? Not applicable
Ethics committee - additional information: This study received Caldicott approval granted by the NHS board.
7 min
Structured reporting quality of chest CT for lung cancer staging: a double cohort study involving radiology residents
Valeria Peruzzi, Udine / Italy
Author Block: V. Peruzzi, L. Cereser, C. Vecchia, F. Cortiula, M. Bortolot, G. Como, R. Girometti, C. Zuiani; Udine/IT
Purpose: To compare radiology residents' (RRs) report quality using the structured radiological model (SRM) from the Royal College of Radiologists with narrative reporting (NR) for chest CT staging of lung cancer. To assess reporting times for NR and SRM.
Methods or Background: A study coordinator preliminarily selected 30 non-small cell lung cancer patients who underwent a baseline staging contrast-enhanced chest CT examination between 2014 and 2022 at our University Hospital. After attending a dedicated training lesson, four third-year RRs (RR1-4) independently reported all the CT examinations in two 2-month-apart separate reading sessions. In the first reading, all the RRs used the NR, while in the second reading, RR1-2 used the NR, and RR3-4 used SRM.
Two chest-devoted radiologists, in consensus, rated the completeness and accuracy of all the NRs and SRMs. Two thoracic oncologists, in consensus, expressed the perceived clarity for the reports from the second reading session. All the quality indicators were expressed on a 100-point scale. The Wilcoxon test was used for statistical analysis.
Results or Findings: Comparing reading sessions, RR3-4 report completeness was significantly higher when using SRM vs. NR (90.7 vs. 74.0, p<0.001), while RR3-4 accuracy and RR1-2 completeness and accuracy values were not significantly different. In the second reading, report completeness, accuracy, and clarity of RR3-4 were significantly higher than RR1-2, with median values of 90.7 vs. 72.8 (p<0.001), 63.1 vs. 58.7 (p=0.04), and 87.3 vs. 68.3 (p<0.001), respectively. Median RR3-4 reporting time was significantly longer than RR1-2 (13.5 min vs. 10.6 min, p<0.001).
Conclusion: The completeness, accuracy, and clarity of SRM were superior to NR at the price of a longer reporting time.
Limitations: Retrospective, monocentric study, with a limited number of patients.
Funding for this study: None
Has your study been approved by an ethics committee? Yes
Ethics committee - additional information: Not applicable

This session will not be streamed, nor will it be available on-demand!