Article Text
PDF
Abstract
The optimal management of small but growing nodules remains unclear. The SUMMIT study nodule management algorithm uses a specific threshold volume of 200 mm3 before referral of growing solid nodules to the multidisciplinary team for further investigation is advised, with growing nodules below this threshold kept under observation within the screening programme. Malignancy risk of growing solid nodules of size >200 mm3 at initial 3-month interval scan was 58.3% at a per-nodule level, compared with 13.3% in growing nodules of size ≤200 mm3 (relative risk 4.4, 95% CI 2.17 to 8.83). The positive predictive value of a combination of nodule growth (defined as percentage volume change of ≥25%), and size >200 mm3 was 65.9% (29/44) at a cancer-per-nodule basis, or 60.5% (23/38) on a cancer-per-participant basis. False negative rate of the protocol was 1.9% (95% CI 0.33% to 9.94%). These findings support the use of a 200 mm3 minimum volume threshold for referral as effective at reducing unnecessary multidisciplinary team referrals for small growing nodules, while maintaining early-stage lung cancer diagnosis.
- lung cancer
- imaging/CT MRI etc
This is an open access article distributed in accordance with the Creative Commons Attribution 4.0 Unported (CC BY 4.0) license, which permits others to copy, redistribute, remix, transform and build upon this work for any purpose, provided the original work is properly cited, a link to the licence is given, and indication of whether changes were made. See: https://creativecommons.org/licenses/by/4.0/.
Statistics from Altmetric.com
Introduction
Indeterminate pulmonary nodules are common in lung cancer screening, with only a small proportion ultimately confirmed as malignant. On baseline scans, malignancy risk and thus nodule management in solid nodules is primarily driven by size1–3 whereas at follow-up CT scan, growth indicates an elevated risk of malignancy.3 4 The question of how to optimally manage growing solid nodules which remain below a size threshold for subsequent investigations is an area of uncertainty.
In LungRads 1.1,1 participants with growing nodules that remain <8 mm are recommended for either CT surveillance or positron-emission tomography (PET)/CT scanning. The British Thoracic Society (BTS) guidelines and the European Position Statement5 stipulate that all nodules initially between 80 or 100 mm3 (respectively) and 300 mm3 (or ≥5 to <10 mm in diameter when volumetry is not possible) which subsequently demonstrate growth with volume doubling time (VDT) <400 days are referred for further definitive management, regardless of size.
The SUMMIT study (NCT03934866) is an observational study in high-risk participants using low-dose CT (LDCT) screening in London. The nodule management protocol used in SUMMIT is based on the BTS guidelines but includes a specific threshold volume of 200 mm3 before referral for growing solid nodules to the multidisciplinary team (MDT) for further investigation is advised. The rationale for this was twofold: first, data from the NELSON study6 found that the development of new solid pulmonary nodules was associated with a higher cancer risk, but only above a threshold volume of 206 mm6. The implication is that nodules below this size, even if growing, have a lower risk of malignancy and do not require definitive investigation at this stage. Second, there are particular challenges when performing further investigations on nodules smaller than 8 mm/200 mm3 as they are typically below the resolution limits of positron-emission tomography (PET)/CT and technically more difficult to biopsy.
There is little previous data from studies that have prospectively managed small growing nodules in this way. The aim of this study was to assess the safety and efficacy of this approach.
Methods
The SUMMIT study is a prospective observational cohort study to examine the performance of delivering a LDCT screening service to a high-risk population in London and to validate a multicancer early detection blood test (ClinicalTrials.gov NCT03934866). Eligible participants were 55–77 years old, met the US Preventive Services Task Force 2013 screening criteria,7 or had a PLC0m2012 risk of ≥ 1.3%8 and attended three annual lung health checks (baseline (Y0), year 1 (Y1) and year 2 (Y2)) with LDCT. Study scans were performed without contrast at maximal inspiration in one continuous craniocaudal acquisition with radiation dose optimised based on body weight. Images were read by thoracic radiologists using computer aided detection (CADe) software (Veolity V.1.4, MeVIS Medical Solutions, Germany) and semiautomated volumetry.
The SUMMIT nodule management protocol has been published.9 In brief, solid nodules of ≥80 mm3 and <300 mm3 on baseline scan undergo interval scan at 3 months. Nodule growth was defined as percentage volume change (PVC) of ≥25%, or on visual assessment if baseline volumetry was unreliable. At follow-up, scan nodule stability was defined as PVC of between −24 and+24%, or stable diameter in cases of unreliable segmentation. Nodules stable at 3-month interval scan undergo further scans at Y1 and Y2. Nodules demonstrating growth with volume >200 mm3 are referred to the MDT for definitive assessment. Growing nodules which remain ≤200 mm3 are scheduled to undergo a further surveillance scan after another 3-month interval and are only referred for MDT assessment when volume exceeds 200 mm3. At all time points, protocol deviation was permissible based on radiologists’ assessment in individual cases.
This study analyses outcomes from all participants who had a baseline scan between study commencement (4 April 2019) and temporary closure for the Sars-CoV-2 pandemic on 18 March 2020 (N=11 566), with solid nodules between 30 and 300 mm3 at baseline CT, showing growth at a 3-month interval scan. Nodules interpreted as benign intrapulmonary lymph nodes at baseline did not undergo surveillance and are excluded from this analysis.
Solid nodules present on baseline scan but not marked on initial review, and subsequently noted to be growing on 3-month interval scan (referred to as ‘retronodules’) were included in this analysis. Growing nodules which had unreliable volumetry at follow-up scan were excluded from this analysis. Participants who had their scheduled 3-month interval scan delayed beyond 6 months from baseline (primarily due to the SARS-CoV2 pandemic) were also excluded from this analysis.
Cancer was confirmed by histology or diagnosed clinico-radiologically by MDT assessment. Nodules were recorded as benign based on any of the following criteria: (1) benign histology following MDT referral; (2) volume stability or decrease in volume over at least 12 months (PVC<25% and/or volume-doubling time (VDT)>600 days2); (3) stability on 2D measurements over 2 years where volumetry was unreliable or (4) resolution of nodules.
Measured outcomes were (1) false-negative rate, defined as the proportion of all growing nodules managed initially by surveillance and subsequently diagnosed as lung cancer at greater than stage 110; (2) positive predictive value (PPV) of our protocol, defined as [(all cancers diagnosed) ÷ (all growing nodules >200 mm3)] and (3) Relative risk of malignancy at 2 years, defined as the percentage of nodules proved malignant within 2 years in those that demonstrated growth and had volume >200 mm3 at initial 3-month interval divided by the percentage of nodules proved malignant within 2 years in those that demonstrated growth but remained ≤200 mm3 at initial 3-month interval scan. Fishers’ exact test was used to assess proportional differences with statistical significance defined as p value of less than 0.05.
Analysis was performed using R Studio (V.4.0).
Results
Of the 11 566 participants who underwent a baseline scan in this analysis, 3621 solid nodules with volume 30–300 mm3 were identified in 2344 participants (figure 1). Of these, 1620 nodules in 953 participants underwent a 3-month interval scan (nodules 30–80 mm3 do not undergo interval scanning in BTS/SUMMIT protocol so these scans were performed for a coexisting finding (larger nodule or consolidation)). At initial interval scan, 1424/1620 (88%) nodules persisted, while the remaining 196 (12%) resolved. Seventy-nine nodules in 70 participants demonstrated growth at 3-month scan, with an additional two nodules in two participants showing clear growth on visual assessment but with unreliable segmentation at follow-up CT; these were excluded from further analysis. A further 17 nodules in 17 participants were noted to be growing on 3-month interval scan having been initially missed or disregarded on baseline scan giving a total of 96 solid nodules in 87 participants with clear evidence of growth at the 3-month scan. Of the 96 growing solid nodules, 36 nodules in 30 participants had volume >200 mm3 (management and outcomes shown in figure 2); and 60 nodules in 59 participants had volume ≤200 mm3 (management and outcomes shown in figure 3).
CONSORT diagram for participants in this analysis. Numbers in italics are nodule volume range (median) in mm3. Participant numbers add up to greater than total due to participants having multiple nodules in different categories +96 growing nodules includes 79 nodules noted at baseline and 17 nodules present but not marked on initial scan and subsequently seen to be growing (‘retronodules’) Ppt, participant.
Outcomes of growing solid nodules >200 mm3 at first interval scan. Numbers given at per-nodule and per-participant level. Numbers in italics are nodule volume range and (median), mm3. AProtocol deviation due to radiologist assessment: interpreted as intrapulmonary lymph nodes (n=2) or likely inflammatory (n=5). BDefined as VDT >600 days. CRadiologists’ decision based on morphology. MDT, multidisciplinary team; ppt, participant; VDT, volume doubling time.
Outcomes of growing solid nodules ≤200 mm3 at first interval scan. Numbers given at per-nodule and per-participant level. An=2 growing solid nodule >200 mm3 covered in figure 2, n=1 lymph node mass, n=3 growing part-solid nodules. BProtocol deviation due to radiologist interpretation: (likely inflammatory n=2, IPLNs n=7). CNo thoracic cancer diagnosis. DNon-lung cancer cause of mortality. EVDT >600 days. FProtocol deviation due to radiologist interpretation (benign morphological appearances n=2, No growth since 3-month scan n=4). GSmall cell lung cancer separate to nodule under surveillance. IPLN, intrapulmonary lymph node; MPT, multidisciplinary team; ppt, participant; VDT, volume doubling time.
On a per-nodule level, of the 96 growing solid nodules included in this analysis, 29 nodules (30.2%) were ultimately identified as malignant. Of these, 22/29 (75.9%) nodules were bronchogenic carcinomas in 22 participants, comprising 18 lung cancers diagnosed on histology and four lung cancers diagnosed clinicoradiologically by the MDT (based on CT and PET/CT findings due to patient preference or fitness). The remaining 7/29 nodules in one patient were clinicoradiologically diagnosed lung metastases from a subsequently diagnosed primary oesophageal cancer.
Of the 36 nodules showing growth with volume>200 mm3 at first interval scan (figure 2), 21/36 (58.3%) were malignant in 15/30 (50%) participants (table 1). In the 60 nodules ≤200 mm3 at initial 3-month interval scan (figure 3), the risk of a growing nodule being diagnosed as cancer within 2 years was 13.3% on a per-nodule (8/60) and 13.6% on per-participant (8/59) basis (relative risk of nodule malignancy 4.4, (95% CI 2.17 to 8.83), two nodules (two participants) ≤200 mm3 at initial 3-month interval scan underwent further surveillance beyond 2 years).
Volume at first interval scan and probability of malignancy of growing solid nodules
The PPV of a growing nodule that reached size >200 mm3 being malignant was 65.9% (29/44), constituting 21 cancers in 36 nodules >200 mm3 at first interval scan (figure 2) plus 8 cancers in 8 nodules which grew to >200 mm3 at subsequent scans (figure 3). PPV was 60.5% (23/38) on a per-participant basis.
In this study, 53 growing nodules of ≤200 mm3 at 3 months were managed by further surveillance (figure 3). Of these, eight were diagnosed as lung cancer within 2 years. Of these, seven were stage 1 and one was stage 3 (pN2 nodal involvement at surgical resection, not demonstrable on preoperative CT imaging). The false negative rate of the protocol was therefore 1.9% (95% CI 0.33% to 9.94%) (1/53).
At first interval scan, median VDT was shorter in nodules subsequently confirmed to be malignant compared with those where malignancy was ultimately excluded (median 98 (range 42–389) days vs median 202 (range 27–440) days). At first interval scan, a nodule management protocol based on evidence of growth alone would have resulted in all 87 participants being referred for definitive assessment2; a combination of growth and minimum volume threshold reduced this by 62% to 33/87 (figures 2 and 3). Example images are shown in figure 4.
Panel A: Growing nodule subsequently diagnosed as lung cancer. (A) Baseline scan, volume 42 mm3 (B) 3 months, volume 92 mm3, PVC +117%, VDT 98 days. (C) 6-month scan (performed at 8 months), volume 246 mm3, PVC +168%, VDT 109 days (referred at this time). Panel B: Growth seen at first interval scan, subsequently stable over 2 years. (A) Baseline, volume 82 mm3 (B) 3 months, volume 126 mm3, PVC+53%, VDT 153 days. (C) 6-month scan (performed at 8 months), volume 74 mm3, PVC −41%. (D) 12 months, volume 53 mm3, PVC -28% (E) 24 months, volume 58 mm3. PVC, percentage volume change; VDT, volume doubling time.
Discussion
In the NELSON study, the malignancy risk of new nodules at interval scan<206 mm3 was 3.1%.6 Our data show the malignancy risk in nodules which grow on first interval scan (performed at 3–6 months after baseline) but remain ≤200 mm3 was over fourfold higher, at 13.3%.
However, our results provide support for a conservative approach involving close CT observation in growing nodules ≤200 mm3. Of the nodules in this category that were malignant, all but one remained at stage I, with an overall false-negative rate of 1.9%.
Importantly, this approach avoids unnecessary MDT referral and further investigation for a finding which is benign or indolent in 86% of cases. By contrast, our results indicate that growing nodules >200 mm3 require further investigation for lung cancer, with a PPV for malignancy of 65.9% on a per-nodule or 60.5% on a per-participant basis. This approach has two potential benefits: it focuses MDT discussion on cases that are more likely to represent lung cancer and which have reached a size for meaningful intervention, while ensuring that smaller nodules can remain within the protocol-driven, streamlined management and safety netting provided by a screening programme. We anticipate that this strategy would reduce both variability in and overall rates of PET-CT referral at a stage when such nodules would be too small to evaluate.
Key strengths of our study include our large cohort size and that our management approach to nodules ≤200 mm3 was implemented prospectively. Furthermore, all studies were read by experienced thoracic radiologists with standardised scanners and CADe software ensuring consistency.
A limitation of our study is that as our protocol used a 200 mm3 volume threshold, only nodules with reliable segmentation at follow-up scan were included in this analysis. While this allows us to validate this approach (and volumetric assessment is currently recommended by national2 and European5 guidelines), it means a small number of nodules where reliable volumetric analysis could not be achieved were excluded. This may limit generalisability in contexts where nodule volumetry is not routinely available or for nodules where volume cannot be accurately assessed. Furthermore, although the SUMMIT study includes a large number of participants, as growing solid nodules comprise only a small proportion of total screen-detected nodules, focusing specifically on this group limits the number of nodules and cancers in this final analysis. Our findings should therefore be taken cautiously, and this approach would benefit from prospective validation in further cohorts. Finally, it is important to recognise that this approach was used within a screening context, where participants would return for further annual or biennial scans to identify more slowly growing nodules. Nevertheless, there is precedent for nodule management approaches derived from screening programmes to be used in guidelines for the management of incidentally detected nodules, including the Brock score and volume-doubling time.
In conclusion, we provide unique, prospective evidence that a solid nodule management protocol encompassing a combination of growth and minimum size threshold is safe and reduces unnecessary MDT referrals for benign lesions, while maintaining early-stage lung cancer diagnosis.
Ethics statements
Patient consent for publication
Ethics approval
This study involves human participants and was approved by London—City and East Research Ethics Committe17/LO/2004. Participants gave informed consent to participate in the study before taking part.
Footnotes
AD and SMJ are joint first authors.
Twitter @LUNGRADIOLOGIST
Collaborators Dickson Thomas Callender¹, Mamta Ruparel¹, Kitty Chan², Rachael Sarpong², Malavika Suresh², Samantha L Quaife³, Vicky Bowyer⁴, Ethaar El-Emir ⁴, Judy Airebamen⁴, Alice Cotton⁴, Kaylene Phua⁴, Elodie Murali⁴, Simranjit Mehta⁴, Janine Zylstra⁴, Karen Parry-Billings⁴, Columbus Ife⁴, April Neville⁴, Paul Robinson⁴, Laura Green⁴, Zahra Hanif⁴, Helen Kiconco⁴, Ricardo McEwen⁴, Dominique Arancon⁴, Nicholas Beech⁴, Derya Ovayolu⁴, Christine Hosein⁴, Sylvia Patricia Enes⁴, Qin April Neville⁴, Jane Rowlands⁴, Aashna Samson⁴, Urja Patel⁴, Fahmida Hoque⁴, Hina Pervez⁴, Sofia Nnorom⁴, Moksud Miah⁴, Julian McKee⁴, Mark Clark⁴, Jeannie Eng⁴, Fanta Bojang⁴, Claire Levermore⁴, Anant Patel⁷, Sara Lock⁸, Rajesh Banka⁹, Angshu Bhowmik¹⁰, Ugo Ekeowa¹¹, Zaheer Mangera¹², William M Ricketts¹³, Neal Navani⁴, Terry O’Shaughnessy¹³, Charlotte Cash⁷, Magali Taylor⁴, Samanjit Hare⁷, Tunku Aziz¹³, Stephen Ellis¹³, Anthony Edey¹⁴, Graham Robinson¹⁵, Alberto Villanueva¹⁶, Hasti Robbie¹⁷, Elena Stefan¹⁸, Charlie Sayer¹⁹, Nick Screaton²⁰, Navinah Nundlall⁴, Lyndsey Gallagher⁴, Andrew Crossingham⁴, Thea Buchan⁴, Tanita Limani⁴, Kate Gowers¹, Kate Davies¹, John McCabe¹, Joseph Jacob¹,²⁴, Karen Sennett²¹, Tania Anastasiadis ²², Andrew Perugia²³, James Rusius²³.
Contributors All authors were involved in the design and conduct of the SUMMIT study. AWC drafted the manuscript and analysed data with input from AD, AN and SMJ. AD and SMJ are joint last authors. All authors contributed to and approved the final manuscript.
Funding The SUMMIT study is funded by GRAIL through a research grant awarded to SMJ as principal investigator. SMJ was a Wellcome Trust Senior Fellow in Clinical Science (WT107963AIA). SMJ is supported by CRUK programme grant EDDCPGM\100002, the Rosetrees Trust, the Roy Castle Lung Cancer foundation, the Garfield Weston Trust and UCLH Charitable Foundation.
Competing interests AWC, CH, JLD, ST, HH, PV, RP and AB are all funded or part-funded through GRAIL as part of the SUMMIT Study. SUMMIT is sponsored and conducted by University College London and funded by GRAIL LLC through a research grant awarded to SMJ as principal investigator. SMJ’s full disclosures are as a Paid Advisory Board member Astra-Zeneca, Bard1 Bioscience, Achilles Therapeutics, Jansen. Assistance for travel to meetings from Astra Zeneca, Takeda, and grant income from GRAIL Inc, Owlstone and share options from Optellum; BARD1 Lifescience. AN is part-funded through the UCLH Biomedical Research Centre. AD’s disclosures are personal fees from Boehringer Ingelheim, Roche, Galacto Biotech, Galapagos, Brainomix and Vicore. AH’s disclosures are consulting fees to Evidera and assistance for travel to meetings from GRAIL.
Provenance and peer review Not commissioned; externally peer reviewed.