Research Presentation Session

RPS 1404a - Implementing lung cancer screening

Lectures

1
RPS 1404a - Identifying participant subgroups in a lung cancer CT screening setting based on competing risks at the baseline scan

RPS 1404a - Identifying participant subgroups in a lung cancer CT screening setting based on competing risks at the baseline scan

06:17A. Schreuder, Nijmegen / NL

Purpose:

Cardiovascular disease (CVD) and COPD are the main competing causes of death which diminish the benefits of lung cancer (LC) CT screening. Being able to identify participants at risk may encourage a multidisciplinary preventative approach towards reducing overall mortality.

Methods and materials:

The model derivation (n=23,096) and validation cohorts (n=2,287) were formed using data from the NLST and MILD trials, respectively. CT measures of CVD and COPD were extracted using computer algorithms, and nodule features were provided by the trials. Risk models were developed to predict four outcomes: LC incidence, LC mortality, CVD mortality, and COPD mortality. For each outcome, a Cox regression model was derived from combining patient characteristics and CT features from the baseline scans.

Results:

In the NLST, 756 were diagnosed with LC (3.4%) and 259 (1.2%), 435 (2.0%), and 0.5% died of LC, CVD, and COPD, respectively, after five years’ follow-up. 34.2% (259/756) of the LC patients died. Among those who also had a CVD and COPD mortality risk above the 0.9 quantile, this proportion increased to 48.4% (45/93) (Figure 1). In participants with a low LC risk and high CVD and COPD mortality risks, respectively, 0.8% were diagnosed with LC (3/362) and 7.5% died (27/362) (“high” and “low” risk defined in Figure 2). Among those with a high LC risk and low CVD and COPD mortality risks, 16.0% were diagnosed with LC (150/940) and 2.1% died (20/940).

Conclusion:

LC screening participants can be stratified into groups based on LC, CVD, and COPD risk and may benefit from personalised follow-up protocols. Participants with a high LC risk and low risk of CVD or COPD deaths may benefit most from screening.

Limitations:

A retrospective design. Recalibration required.

Ethics committee approval

n/a

Funding:

No funding was received for this work.

2
RPS 1404a - Pulmonary nodule growth: can follow-up be shortened with a high-end or an ultra-high-resolution CT scanner?

RPS 1404a - Pulmonary nodule growth: can follow-up be shortened with a high-end or an ultra-high-resolution CT scanner?

05:56D. Grob, Nijmegen / NL

Purpose:

To determine the interscan variability of pulmonary nodule volume measurements in CT scans acquired with state-of-the-art wide-area and ultra-high-resolution CT systems.

Methods and materials:

In this prospective study, patients with at least two non-calcified solid pulmonary nodules suspicious for metastases on previous CT scans were imaged twice with either a high-end 320‑detector CT (MDCT, Aquilion ONE Genesis, Canon, slice thickness 0.5 mm, 512x512 matrix) or an ultra-high-resolution CT (UHRCT, Precision, Canon, 0.25 mm, 1024x1024). In between scans, an off‑and‑on table strategy was used to simulate follow-up scans with no nodule growth. Semi‑automated volumetric nodule segmentation and volume estimation (max. 4 per patient, effective diameter 4-15 mm) were performed on a lung screening workstation (Veolity). 95%-limits of agreement (LOA) and the time to estimate actual nodule growth rate at a nodule volume doubling time (VDT) of 400 days were calculated.

Results:

17 patients (60 nodules, average volume: 218 mm3) were imaged on the MDCT and 27 patients (90 nodules, 177 mm3) on the UHRCT at a similar dose (mean dose-length-product: 126.6 mGycm vs. 127.2 mGycm, respectively (p=0.98)). The 95%-LOA was ±7.0% for the MDCT and ±5.9% for the UHRCT (p=0.07). Therefore, the minimum required interscan period to detect a VDT of 400 days is only 33-39 days.

Conclusion:

Both scanners result in low interscan variability, especially compared to current clinical standards, which requires a volume change of 25% (the current 95%-LOA) as significant nodule growth. Therefore, the follow-up period to detect pulmonary nodule growth could be dramatically shortened from three to about one month, reducing patient anxiety and the potential for stage shift in lung nodule management.

Limitations:

Pulmonary metastases instead of incidental nodules were measured.

Ethics committee approval

Review board approval and written informed consent were obtained.

Funding:

Canon provided financial funding.

3
RPS 1404a - Volume doubling times of lung adenocarcinomas: correlation with predominant histologic subtypes and prognosis

RPS 1404a - Volume doubling times of lung adenocarcinomas: correlation with predominant histologic subtypes and prognosis

06:04S. Park, Seoul / KR

Purpose:

To investigate differences in VDT between the predominant histologic subtypes of lung adenocarcinomas and to assess the correlation between VDT and prognosis.

Methods and materials:

This retrospective study included patients who underwent at least two serial CT scans before surgery between July 2010 and December 2018. Three-dimensional tumour segmentation was performed on two CT scans and VDTs were calculated. VDTs were compared between predominant histologic subtypes and lesion types using Kruskal–Wallis tests. Disease-free survival (DFS) was obtained from patients undergoing surgery before July 2017. A Cox proportional hazards model was used to determine predictors of DFS.

Results:

Out of 268 patients (mean age, 64 years; 143 men), there were 30 lepidic, 87 acinar, 109 papillary, and 42 solid/micropapillary predominant subtypes. The median VDT was 528.8 days (range, 30.9–2,911.4 days) for all lung adenocarcinomas. VDTs differed significantly across subtypes (P<0.001), being shortest in solid/micropapillary subtypes (229.2 days). In terms of lesion type, solid lesions had significantly shorter VDTs than subsolid lesions (P<0.001).

In the 148 patients included in the survival analysis, 35 had disease recurrence or died. The multivariate analysis identified VDT (≥400 or <400 days) as an independent risk factor for DFS (hazard ratio, 2.310; P=0.024), as well as TNM stage and lesion type. Adding VDT to TNM stage and lesion type led to significant improvement in model performance (C-index, 0.767 vs. 0.795; P=0.037).

Conclusion:

VDTs varied significantly according to the predominant histologic subtypes of lung adenocarcinoma, and VDT had additional prognostic value for DFS.

Limitations:

Retrospective in nature. A selection bias.

Ethics committee approval

Informed patient consent was waived by the institutional review board.

Funding:

Funding from the Basic Science Research Program through the National Research Foundation of Korea funded by the Ministry of Science, ICT & Future Planning.

4
RPS 1404a - Variation in LungRADS™ scoring for screen detected lung nodules

RPS 1404a - Variation in LungRADS™ scoring for screen detected lung nodules

06:59H. Schmidt, Toronto / CA

Purpose:

To determine the Canadian feasibility of a provincial scale roll-out of an organised lung cancer screening program, Cancer Care Ontario (CCO) implemented the multicentre High Risk Lung Cancer Screening Pilot (HRLCSP) in 2017. One of the integral components of the HRLCSP is radiology quality assurance (QA) to support consistency in performing and interpreting screening scans. Participating radiologists were required to attend a workshop, including a pre-workshop quiz.

Methods and materials:

Before the HRLCSP workshop, 43 participating radiologists (chest radiologist without specific experience in lung cancer screening) received the LungRADS™ classification chart and a PowerPoint presentation with 20 questions regarding the scoring and follow-up of 16 different examples of CT images with screen-detected nodules. They were asked to score the nodule examples according to the LungRADS™ classification system. Responses were compared to the consensus of expert radiologists, which served as the reference standard.

Results:

The LungRADS™ scores assigned by the participating radiologists varied significantly and were different from the reference standard. For the 20 questions, the correct answers ranged from 42% to 93%, with a mean of 74%. Nodule scores were both overestimated as well as underestimated.

Conclusion:

Scoring screen-detected nodules using LungRADS™ does vary significantly and radiologist training before reading for a screening program is important for quality and consistency. The application of a pre-workshop quiz allows for the tailoring of the workshop content to address specific knowledge gaps.

Limitations:

n/a

Ethics committee approval

n/a

Funding:

No funding was received for this work.

5
RPS 1404a - Comparative cost-effectiveness of dynamic contrast-enhanced computed tomography versus positron emission tomography in the characterisation of solitary pulmonary nodules: the SPUtNIk trial

RPS 1404a - Comparative cost-effectiveness of dynamic contrast-enhanced computed tomography versus positron emission tomography in the characterisation of solitary pulmonary nodules: the SPUtNIk trial

06:25J. Weir-Mccall, Cambridge / UK

Purpose:

To compare the cost-effectiveness of dynamic contrast-enhanced computed tomography (DCE-CT) and 18Fluorine-Fluorodeoxyglucose positron emission tomography/computed tomography (PET-CT) based approaches for the diagnosis of solitary pulmonary nodules (SPN).

Methods and materials:

In this prospective multicentre trial, 380 participants with indeterminate SPN (8-30 mm) underwent DCE-CT and PET-CT. All patients completed 2 years follow-up with a collection of subsequent investigations, management strategy, and outcome. The cancer prevalence, diagnostic accuracy, and test outcome results of the SPUtNIk trial were used in an economic evaluation decision model. Further data on complications of the tests and survival came from a systematic review and meta-analysis. NHS tariff costs and the literature were used for cost estimates. The economic evaluation was reported as a cost-consequence analysis and the incremental cost per correctly managed case over a two-year time horizon was also calculated.

Results:

Sequential DCE-CT and PET-CT were most accurate (84.4±1.4%) for the management of SPN compared with DCE-CT (77.8±2%) or PET-CT alone (82±1.6%). However, the combined modality approach was most costly in the cost-consequence analysis (£4058±210). A DCE-CT strategy was the least costly (cost=£3305±199) followed by PET/CT (£4014±206). For all correctly managed cases, the ICER for DCE-CT/PET-CT over DCE-CT was £11,323. For DCE-CT/PET-CT versus DCE-CT, the incremental cost per correctly managed case was £11,323. PET/CT would not be cost-effective compared to either DCE-CT or DCE-CT/PET-CT.

Conclusion:

If society is not willing to pay more than £11,323 per accurately managed case, DCE-CT is a cost-effective strategy. Above this value, DCE-CT/PET-CT is the most cost-effective strategy. With the greater availability and access to CT, and lower cost, consideration should be given to using DCE-CT as the first-line test for indeterminate pulmonary nodules.

Limitations:

Limited availability of utility data and long-term costs.

Ethics committee approval

REC-12/SW/0206.

Funding:

NIHR-HTA (grant no:09/22/117).

6
RPS 1404a - Quality assurance in lung cancer screening by computed tomography: optimised ultra-low radiation dose by beam filtering in a randomised study

RPS 1404a - Quality assurance in lung cancer screening by computed tomography: optimised ultra-low radiation dose by beam filtering in a randomised study

05:00M. Silva, Parma / IT

Purpose:

To test different protocols of ultra-low radiation dose computed tomography (ULDCT) by x-ray spectrum shaping for lung cancer screening (LCS).

Methods and materials:

Based on a former anthropomorphic phantom study, 375 subjects were prospectively randomised to one of the following ULDCT protocols: fully automated exposure control (both voltage and current - “ULDCT1”), fixed tube voltage (Sn100kVp) and current according to patient size (140mAs or 210mAs – “ULDCT2”), and a hybrid approach with fixed tube voltage and automatic exposure control for current (“ULDCT3” Sn100-100refmAs and “ULDCT4” Sn150-20refmAs). Each subject underwent double CT acquisition with one of the above protocols and the low-dose CT (LDCT: 120kVp, 25mAs). Intrasubject CT dose index (CTDIvol) and dose-length product (DLP) were compared (mean and standard deviation reported). An analysis of variance to compare ULDCT protocols was undertaken.

Results:

CTDIvol was 1.62 (0) mGy in LDCT compared to ULDCT: 0.59 (0.21), 0.72 (0.15), 0.51 (0.16), and 0.66 (0.17).

DLP was 65 (6) mGycm in LDCT compared to ULDCT: 24 (8), 29 (7), 21 (7), 27 (8).

ULDCT3 showed lower CTDIvol than any ULDCT protocol (p<0.001) and lower DLP than ULDCT2 and 4 (p<0.001). Conversely, ULDCT2 appeared the less convenient approach for dose saving, with CTDIvol and DLP higher than ULDCT1 and 3 (p<0.001).

Conclusion:

Optimised ULDCT protocols with x-ray spectrum shaping by tin-filter can be used with fixed tube voltage at Sn100 kVp and automatic control of tube current (ULDCT3) to aim at the sharpest optimisation of radiation dose. Fixed techniques are less effective in reducing doses.

Limitations:

Image quality will be tested for consistency of each optimised protocol within LCS workflow by dedicated post-processing platform (detection and semi-automatic volume of nodule).

Ethics committee approval

Prot INT 21/11 and further amendments.

Funding:

Italian Ministry of Health, AIRC, Fondazione Cariplo, Lombardy Region.

7
RPS 1404a - Using spirometry to identify high-risk individuals not eligible for lung cancer screening

RPS 1404a - Using spirometry to identify high-risk individuals not eligible for lung cancer screening

06:16J. Kavanagh, Toronto / CA

Purpose:

To determine if FEV1 can identify a high-risk subgroup of individuals traditionally excluded by current lung cancer screening eligibility criteria.

Methods and materials:

Spirometry (2009-2013) was administered in a lung screening program and patient outcomes were tracked until 2017. The addition of forced-expiratory-volumes-at-one second, FEV1, to four different screening-eligibility criteria, (DLST, NLST, NELSON, and OLSP) was retrospectively assessed to identify high-risk individuals. Sensitivity, specificity, and the number needed to screen at various FEV1 cutoffs were compared to determine the optimal clinical cutoff.

Results:

In 1,161 participants, optimal clinical cutoffs for FEV1 were 90%, 90%, 90%, and 85% for the DLST, NLST, NELSON, and OLSP, respectively. Respectively, the sensitivity was increased significantly by 33%, 28%, 28%, 16%, specificity was reduced by 19%, 25%, 24%, 22%, and NNS was reduced by 122, 7, 10, -3 when compared to original criteria alone.

Conclusion:

Spirometry can identify a high-risk subpopulation missed by original lung screening criteria, particularly when self-reported COPD is not included. Inclusion of spirometry-identified airway obstruction screening-eligibility criteria into screening programs that lack another method for COPD assessment increases sensitivity.

Limitations:

A retrospective single-institution study. The FEV1 cutoffs were chosen by the authors pragmatically to maximise sensitivity as well as keeping specificity acceptable.

Ethics committee approval

Approval was obtained through the University Health Network (UHN) Research Ethics Board (REB 06-0639). Recruitment began in June 2003 and concluded in December 2009 as part of the Lusi Wong Princess Margaret Early Lung Cancer Detection Program (PM-ELCAD).

Funding:

This study was supported by the Lusi Wong Fund, the Princess Margaret Cancer Foundation, the Alan Brown Chair in Molecular Genomics (to GL), and the University of Toronto Comprehensive Research Experience for Medical Students programme.

8
RPS 1404a - The classification of pulmonary nodules by Lung-RADS 1.1: a randomised prospective analysis of four ultra-low dose CT protocols in a lung cancer screening trial

RPS 1404a - The classification of pulmonary nodules by Lung-RADS 1.1: a randomised prospective analysis of four ultra-low dose CT protocols in a lung cancer screening trial

04:54G. Milanese, Parma / IT

Purpose:

To analyse the intrasubject agreement between low-dose (LDCT) and ultra-low-dose CT (ULDCT) for LungRADS1.1 in a lung cancer screening (LCS) setting.

Methods and materials:

375 subjects underwent single-breath-hold double CT acquisition protocol, including both standard LDCT scan (120kVp, 25mAs) and one of the following ULDCT protocols: fully automated exposure control (“ULDCT1”), fixed tube-voltage and current according to patient size (“ULDCT2”: Sn100kVp, 140mAs/210mAs), or hybrid approach with fixed tube-voltage and automated exposure control for current (“ULDCT3”: Sn100kVp-100mAsref; “ULDCT4”: Sn150kVp-20mAsref).

Prospective LDCT reading was performed by semi-automatic software for the detection and segmentation of pulmonary nodules (PN). After a wash-out time of at least 2 weeks, two independent readers repeated the same process on ULDCT, without manual corrections. Dominant PNs (dPNs) were selected from both LDCT and ULDCT according to volumetric LungRADS1.1.

Agreement was assessed by means of Cohen’s weighted-kappa test.

Results:

Intrasubject inter-scanning protocol variation for LungRADS1.1 was: κULDCT1=0.8461, κULDCT2=0.9030, κULDCT3=0.8813, and κULDCT4=0.9033 for Reader1, and κULDCT1=0.8469, κULDCT2=0.9029, κULDCT3=0.7896, and κULDCT4=0.7844 for Reader2.

Inter-reader agreement by ULDCT was: κULDCT1=0.8726, κULDCT2=0.8899, κULDCT3=0.8466, and κULDCT4=0.8590.

Reader1 missed dPNs in 11/98 (11.2%) ULDCT1, 12/94 (12.8%) ULDCT2, 16/88 (18.1%) ULDCT3, and 14/95 (14.7%) ULDCT4. Missing of LDCT dPNs did not affect management by ULDCT3.

Reader2 missed dPNs in 13/98 (12.2%) ULDCT1, 9/94 (9.6%) ULDCT2, 11/88 (12.5%) ULDCT3, and 8/95 (8.4%) ULDCT4. Missing of LDCT dPNs would have affected management by ULDCT3 for 1 subject (shift from LungRADS3 to LungRADS2).

Conclusion:

Intrasubject and inter-reader agreement for LungRADS between LDCT and ULDCT ranged from good to very good. ULDCT3 scanning protocol can be proposed for the purposes of LCS.

Limitations:

Single CT and CAD used.

Ethics committee approval

Prot-INT21/11 and further amendments.

Funding:

Italian Ministry of Health, AIRC, Fondazione Cariplo, Lombardy Region.

9
RPS 1404a - Lung-RADS category and smoking status can predict adherence to recommendations in a real-world low-dose CT lung cancer screening program

RPS 1404a - Lung-RADS category and smoking status can predict adherence to recommendations in a real-world low-dose CT lung cancer screening program

05:51E. Barbosa Jr., Philadelphia / US

Purpose:

Low-dose CT lung cancer screening (LCS) can decrease lung cancer related mortality in persons with a significant smoking history, however, it is hindered to low adherence to follow up recommendations.

Methods and materials:

We retrospectively enrolled all persons who underwent multiple (2 or more) LCS exams from 2014 to 2019 at our institution and recorded patient demographics, lung-RADS category, outcomes, and adherence to screening recommendations. We assessed predictors of adherence via univariate and multivariate logistic regression (MLR).

Results:

260 persons returned for follow-up scans (57.7% had 2, 34.2% had 3, 7.7% had 4, and 0.4% had 5). 43/260 (16.5%) had positive scans, of which 28/260 (10.8%) were lung-RADS category 3, 8/260 (3.1%) were 4A, 6/260 (2.3%) were 4B, and 2/260 (0.8%) were 4X. 143 persons were current smokers at the time of the baseline LDCT and 117 persons were former smokers who quit within the last 15 years before the baseline LDCT. Overall adherence was 43.0% but increased progressively with higher lung-RADS (33.2%-100%) and with former smoker status (50.0%, vs 36.2% for current smokers). Smoking status and a positive lung-RADS category were the only statistically significant predictors of adherence in MLR modelling.

Conclusion:

Adherence to LCS recommendations, which is crucial to maximize the benefits of LCS and allow early diagnosis of lung cancer, is less than 50% and is lower in persons who are current smokers and with negative LCS exams, offering a roadmap for targeted performance improvement, which healthcare systems can leverage to improve LCS cost-effectiveness and maximize its societal benefits.

Limitations:

A retrospective design.

Ethics committee approval

IRB approved.

Funding:

No funding was received for this work.

PEP Subscription Required

This course is only accessible for ESR Premium Education Package subscribers.