Abstract
Background: Lung cancer diagnosis, staging and treatment may be enhanced by multidisciplinary participation and presentation in multidisciplinary meetings (MDM). We performed a systematic review and meta-analysis to explore literature evidence of clinical impacts of MDM exposure.
Methods: A study protocol was registered (PROSPERO identifier CRD42021258069). Randomised controlled trials and observational cohort studies including adults with nonsmall cell lung cancer and who underwent MDM review, compared to no MDM, were included. MEDLINE, CENTRAL, Embase and ClinicalTrials.gov were searched on 31 May 2021. Studies were screened and extracted by two reviewers. Outcomes included time to diagnosis and treatment, histological confirmation, receipt of treatments, clinical trial participation, survival and quality of life. Risk of bias was assessed using the ROBINS-I (Risk of Bias in Non-randomised Studies – of Interventions) tool.
Results: 2947 citations were identified, and 20 studies were included. MDM presentation significantly increased histological confirmation of diagnosis (OR 3.01, 95% CI 2.30–3.95; p<0.00001) and availability of clinical staging (OR 2.55, 95% CI 1.43–4.56; p=0.002). MDM presentation significantly increased likelihood of receipt of surgery (OR 2.01, 95% CI 1.29–3.12; p=0.002) and reduced the likelihood of receiving no active treatment (OR 0.32, 95% CI 0.21–0.50; p=0.01). MDM presentation was protective of both 1-year survival (OR 3.23, 95% CI 2.85–3.68; p<0.00001) and overall survival (hazard ratio 0.63, 95% CI 0.55–0.72; p<0.00001).
Discussion: MDM presentation was associated with increased likelihood of histological confirmation of diagnosis, documentation of clinical staging and receipt of surgery. Overall and 1-year survival was better in those presented to an MDM, although there was some clinical heterogeneity in participants and interventions delivered. Further research is required to determine the optimal method of MDM presentation, and address barriers to presentation.
Shareable abstract
Multidisciplinary meeting presentation is a relatively low-cost intervention providing substantial benefits in receipt of active treatment and survival. Up to 40% of patients remain unpresented, risking disparity and inequity in treatment and outcomes. https://bit.ly/48yNyxi
Introduction
Lung cancer is a heterogeneous and complex cancer, in which diagnosis, staging and treatment decision-making demands careful consideration of diverse patient, disease and management factors. The rapid evolution of diagnostics and therapeutics and the complex need for multimodality therapies in primary, neoadjuvant and adjuvant roles demands diverse expertise from a range of clinical craft groups to achieve optimal decision-making [1]. The complexity of this evaluation has precipitated recommendations that all patients presenting with lung cancer be evaluated in a multidisciplinary context [2–4].
Multidisciplinary discussion aims to increase the utilisation of evidence-based management, improve treatment access and enhance coordination and communication between health professionals [5, 6]. Some of the benefits attributable to multidisciplinary meeting (MDM) presentation include increased accuracy and completeness of diagnosis and staging [7–9], better adherence to therapeutic guidelines [8, 10], increased [7, 11] and earlier provision of treatment [8], decreased length of hospital admission [12], increased enrolment in clinical trials and increased referrals to palliative care [7, 13].
Multidisciplinary evaluation in lung cancer requires the coordinated and timely collaboration of a diverse array of craft groups. This logistical demand has resulted in diverse approaches to multidisciplinary activity with groups meeting weekly to monthly, in person and virtually, prior to and following diagnosis and or treatment in single institutions and across hospital networks [14].
The use of MDM presentation has been explored in a range of retrospective and prospective studies suggesting patient benefits in accuracy and completeness of clinical staging, increased enrolment in clinical trials, variable effects on survival and impacts on management coordination and unwarranted variation [1]. The confirmation of benefit of MDM is critically important to the formulation of practice guidelines around the need for MDM utilisation.
We performed a systematic review and meta-analysis of existing literature to explore the impacts of MDM presentation on management and outcomes in the most common form of lung cancer, nonsmall cell lung cancer (NSCLC). We sought to answer three questions: 1) what is the impact of MDM management on management processes in NSCLC; 2) what impact does MDM presentation have on receipt of treatment; and 3) what impact does MDM presentation have on survival in NSCLC?
Methods
Protocol and registration
A study protocol was developed and registered in the PROSPERO register of systematic reviews accessible at www.crd.york.ac.uk/prospero/ (CRD42021258069).
Eligibility criteria
Inclusion criteria included all subjects with NSCLC diagnoses, in adults aged >18 years. All cancer stages were included. We included randomised controlled trials and prospective or retrospective observational cohort studies comparing patients who underwent lung cancer MDM (tumour board) discussion with patients who had care not including MDM discussion. We excluded patients with small cell lung cancer and other thoracic malignancies.
Information sources and search strategy
A search strategy was devised (supplementary table S1), and databases searched included MEDLINE; Ovid SP from 1946 to date; Cochrane Central Register of Controlled Trials (CENTRAL), via the Cochrane Register of Studies, all years to date; Embase Ovid SP 1974 to date; US National Institutes of Health Ongoing Trials Register ClinicalTrials.gov; World Health Organization International Clinical Trials Registry Platform (apps.who.int/trialsearch) and PsycInfo. Searches were conducted on 31 May 2021 and updated to 2 January 2024.
Selection process
Citation title and abstracts were reviewed in Covidence (www.covidence.org/) independently by two reviewers (A. Harrison and R. Stirling) using specified eligibility criteria and consensus confirmed by study discussion. Studies selected for full-text review were reviewed independently by two reviewers (A. Harrison and R. Stirling) and consensus achieved after discussion with a third reviewer (H. Barnes).
Data collection process
Data were extracted independently by each assessor (V. Lee and R. Stirling) using a standardised data collection form on Covidence. Details regarding study identification, methods, patient population, interventions and outcomes were collected and consensus between assessors achieved by comparing the two forms and discussing any discrepancies.
Data items
Histological diagnosis was identified when studies reported histological confirmation of lung cancer or histological categorisation of adenocarcinoma, squamous cell cancer, large cell cancer or histology not otherwise specified. Clinical staging was confirmed by reporting of tumour, node, metastasis (TNM) stage summary prior to treatment. Receipt of treatment was confirmed by reporting of receipt of surgery, chemotherapy, radiotherapy or palliative care. Clinical trial participation was confirmed by reporting of clinical trial enrolment. Survival was confirmed by survival fractions in MDM and non-MDM cohorts.
Risk-of-bias assessment
Risk of bias was assessed using the ROBINS-I (Risk of Bias in Non-randomised Studies – of Interventions) tool for nonrandomised studies of interventions [15]. Risk of bias was assessed by two reviewers (V. Lee and J. Taverner) and discrepancies resolved with consensus review. This tool evaluates risk in relevant domains including confounding, selection, deviation from intended interventions, missing data, measurement of outcomes and selection of the reported result. Risk-of-bias assessment results are displayed using the robvis visualisation tool [16].
Data synthesis and statistical analysis
Effect measures for dichotomous outcomes were assessed by reporting odds ratios for difference between intervention (MDM exposed) and control groups, providing a ratio of the probability that a particular event will or will not occur. Odds ratios were calculated using the Mantel–Haenszel method using random-effects analysis modelling due to clinical study heterogeneity. Effect measures for continuous outcome survival measures were reported as hazard ratio (HR) or odds ratio, providing a ratio of the survival probability using an inverse variance model and random-effects modelling due to study heterogeneity. All analyses were conducted using the Review Manager (RevMan version 5.3; The Nordic Cochrane Centre, Copenhagen). Statistical heterogeneity was assessed using the Chi-squared and I2 tests. An I2 value between 50% and 75% was regarded as substantial heterogeneity and an I2 value of ≥75% as considerable heterogeneity. Subgroup analyses were performed by study design, by participant type, and by intervention type, to explore heterogeneity.
Results
Study selection
2998 articles were retrieved from the search of databases in addition to hand-searching of reference lists in relevant articles. After removal of 229 duplicates, 2769 titles and abstracts were screened for inclusion and 99 citations selected for full-text review, of which 20 studies were selected for inclusion (figure 1) [5, 7, 17–33].
Preferred Reporting Items for Systematic Reviews and Meta-Analyses flow diagram summarising search results and study selection. MDM: multidisciplinary meeting. #: some studies were removed for a number of reasons.
Study characteristics
The search strategy provided 20 studies for systematic review and 14 studies for meta-analysis (table 1) [5, 7, 17–34]. Among the included studies were 18 retrospective cohort studies, one qualitative research study and one mixed retrospective and prospective observational cohort study. International representation included Australia, USA, Canada, Taiwan and Scotland. Nine were single-centre and 11 multicentre studies. All studies included participants with NSCLC. Ray et al. [20] excluded those with an unknown clinical stage; Peckham and Mott-Coles [29] and Stevens et al. [34] only included those with stage I or II NSCLC; Hung et al. [25] only included those with stage III NSCLC; Freeman et al. [18] only included those with stage I–III NSCLC; and Bydder et al. [23] and Forrest et al. [24] only included those with inoperable stage III or IV NSCLC. Tamburini et al. [32] only included those who underwent surgery with curative intent; and Pan et al. [28] and Wang et al. [31] only included those who received treatment.
Results of studies included in the meta-analysis
Risk of bias in studies
Risk-of-bias assessment results are displayed using the robvis visualisation tool [16]. The risk of bias was adjudged as moderate for seven studies and serious for 13 studies with the main causes of risk of bias being bias due to confounding, bias due to selection of participants and bias due to missing data (figure 2).
Risk-of-bias assessment.
Results of individual studies and syntheses
Management timeliness
Three studies reported data reflecting management timeliness. Boxer et al. [7] reported that patients in the non-MDM group had a slightly longer mean time from diagnosis to surgery, but had shorter mean time to curative radiotherapy, palliative chemotherapy and palliative care referral, although the difference was only statistically significant for those who received chemotherapy with palliative intent. Freeman et al. [8] reported a significantly reduced time from diagnosis to treatment for the MDM-presented group compared to the non-MDM-presented group (19±8 versus 32±11 days, p<0.0001). Bjegovich-Weidman et al. [22] reported on the establishment of an MDM evaluation process reporting a reduction in the time from diagnosis to initiation of treatment falling from a mean of 24 days to a mean of 18 days following MDM inception. There were insufficient data detail to consistently confirm improvement in diagnosis to treatment timeliness for various treatment modalities. There were no available data to report timeliness interval from referral to diagnosis.
Histological confirmation of diagnosis
Four studies reporting on 13 315 subjects evaluated effects of MDM on histological confirmation of diagnosis, patients finding an increased odds of histological confirmation (OR 3.01, 95% CI 2.30–3.95, p<0.00001; 13 400 participants) with considerable statistical heterogeneity (Chi-squared=10.08, p=0.02, I2=70%) (figure 3a; supplementary analyses, supplement 1) [5, 7, 27, 33].
Forest plot of clinical management outcomes associated with multidisciplinary meeting (MDM) presentation: a) tissue confirmation of diagnosis; b) clinical staging; c) clinical trial enrolment. M–H: Mantel–Haenszel.
Clinical staging
Five studies were available in 32 190 subjects evaluating likelihood of documentation of clinical stage [5, 7, 18–20]. There was an increase in availability of clinical staging for those who were discussed at MDM (OR 2.55, 95% CI 1.43–4.56; p=0.002), with substantial statistical heterogeneity (Chi-squared=244.23, p<0.00001, I2=98%) (figure 3b; supplementary analyses, supplement 2).
Receipt of evidence-based treatment
Receipt of surgery was evaluated in seven studies involving 25 779 subjects [5, 7, 21, 27, 29–31]. There was a significantly increased likelihood for MDM managed patients to undergo surgery (OR 2.01, 95% CI 1.29–3.12; p=0.002), with substantial statistical heterogeneity (Chi-squared=201, p<0.00001, I2=97%) (figure 4a). Eight studies in 38 720 subjects found no significant impact of MDM on receipt of radiotherapy (OR 1.35, 95% CI 0.85–2.17; p=0.21), with substantial statistical heterogeneity (Chi-squared=399.08, p<0.00001, I2=98%) (figure 4b) [5, 7, 21, 23, 24, 29–31]. Seven studies on 38 637 subjects found no significant impact of MDM on receipt of chemotherapy (OR 1.67, 0.98–2.83; p=0.06), with substantial statistical heterogeneity (Chi-squared=423.47, p<0.00001, I2=99%) (figure 4c) [5, 7, 21, 23, 24, 30, 31]. Six studies in 25 408 subjects found no significant impact on palliative care evaluation by MDM (0.93, 95% CI 0.47–1.84; p=0.85), with substantial statistical heterogeneity (Chi-squared=283.62, p<0.00001, I2=98%) (figure 4d) [5, 7, 18, 19, 23, 24] (supplementary analyses, supplement 3). MDM evaluation was associated with a significant reduction in receiving no active treatment (OR 0.32, 95% CI 0.21–0.50; p=0.01), with low statistical heterogeneity (Chi-squared=22.44, p=0.0002, I2=85%), reported in four studies including 8057 subjects (figure 4e) [20, 23, 24, 30].
Forest plot of receipt of treatment outcomes associated with multidisciplinary meeting (MDM) presentation: a) surgery; b) radiotherapy; c) chemotherapy; d) palliative care; e) no active treatment. M–H: Mantel–Haenszel.
Continued.
Clinical trial enrolment
Two studies evaluated impacts of MDM in clinical research participation, finding no significant impact (OR 2.11, 95% CI% 0.91–4.89; p=0.08), with high statistical heterogeneity (Chi-squared=26.59, p<0.00001, I2=96%) (figure 3c; supplementary analyses, supplement 4) [5, 18].
Survival
Presentation in an MDM was associated with better overall survival (using HR 0.63, 95% CI 0.55–0.72; p<0.00001), reported from five studies including 50 246 participants (I2=85%) [5, 19, 20, 28, 30], although a single-site study of 988 patients reported no significant effect (OR 1.0, 95% CI 0.86–1.17) [7] (figure 5).
Survival estimates: a) overall survival; b) 1-year survival. HR: hazard ratio; IV: inverse variance; MDM: multidisciplinary meeting.
Two studies reported increased odds of 1-year survival (OR 3.23, 95% CI 2.85–3.68; p<0.00001), with minor statistical heterogeneity (Chi-squared=0.42, p=0.52, I2 0%) [21, 23].
Emergency department presentation
Wang et al. [31] reported a high emergency department burden provided by cancer patients, noting 0.9% of emergency department presentations by cancer patients, with 7.7% of cancer survivors visiting the emergency department, using some 1.4 emergency department services per year. Using a propensity score matching approach evaluating 8172 consecutive, treated, lung cancer diagnoses reduced emergency department presentation by 11% (OR 0.89, 95% CI 0.80–0.98; p<0.022) compared to those not seen in the MDM [31].
Quality of life
Linford et al. [17] performed a qualitative study, interviewing patients, caregivers and physicians, exploring experiences of traditional models of care compared to multidisciplinary clinic (MDC) models. Physician participants described improved communication (attributed to real-time face-to face discussions with colleagues) and indicated that the MDC improved collegiality and collaborative relationships through a better appreciation of each other's roles and expertise. Patient participants reported enhanced convenience and efficiency, noting that concurrent appointments increased the likelihood of supporting caregiver attendance and noting a lower likelihood of being overwhelmed with information and greater consistency in physician messaging.
Bjegovich-Weidman et al. [22] reported on the development of a MDC model within a large integrated nonprofit health system. Findings included a significant reduction of time from diagnosis to definitive treatment, high patient satisfaction and increased patient retention within the clinic, a 28% increase in care delivered to patients, and improved quality of care confirmed by reduction in duplicate testing and concordance with national guidelines on work up and therapy.
Stevens et al. [34] studied management of early stage I–II NSCLC, finding that 58% of subjects were presented to an MDM, and that MDM subjects were more likely to receive curative-intent treatment and an increased likelihood of discussion of early- than late-stage disease.
Hung et al. [25] reported MDT discussion for just 39.4% of stage III subjects, observing an increase in median survival of 41.2 months for those discussed at MDT and 25.7 months for those not discussed, with increased survival likelihood also seen in those with better performance status and undergoing surgery.
Discussion
Summary of findings
This systematic review and meta-analysis found that MDM presentation was associated with significant increase in the likelihood of tissue confirmation of diagnosis, documentation of clinical staging, receipt of surgery and a reduction in the likelihood of no active treatment, with nonsignificant trends to an increase in likelihood of clinical trial participation and receipt of chemotherapy. Overall survival hazard was significantly reduced for MDM-presented patients, with an increased odds ratio for 1-year survival.
Evidence-based treatment and survival
Previous studies of MDM function have suggested evidence of an increase in guideline-concordant management and potential for reduction in unwarranted practice variation [5, 35]. Pooled evidence from this study provides clear evidence of a survival benefit resulting from MDM presentation [5, 19–21, 23, 28, 30].
Organisational quality in MDM implementation
Healthcare quality assessment involves review of measures reflecting the structure, process and outcome of healthcare activities [36]. Process and outcome measures, as here reported, are highly likely to be dependent on the structural quality of the healthcare system, which describes the physical capacity, systems and organisational characteristics of the healthcare facility where care occurs [37]. In included studies there is variable reporting of structural quality, and assumptions of equivalence may be incorrect.
MDM patient evaluation is likely to be the optimum forum for management decision-making, although the implementation of these decisions may not necessarily be assumed. Recent studies identified 28–37% discordance between MDM decision-making and subsequent treatment implementation, with significantly poorer timeliness and survival outcomes associated with delivery of MDM-discordant treatment [38, 39]. These findings may imply that the measurement of treatment implementation concordance may be a useful measure and explanation for variation in process quality.
The timing of MDM presentation is not clearly reported among the included studies. Rogers et al. [30] reported 51% of patients being presented to an MDM prior to treatment, 9% receiving treatment prior to MDM presentation, 5% not being presented to an MDM until 60 days post-diagnosis and 35% not presented to an MDM. MDM presentation is likely to have impacts on management decision-making both prior to and following definitive management, including assistance with adjuvant therapy decision-making post-resection.
Opportunities for improvement in MDM function
A recent review evaluated quality of care decisions made by multidisciplinary cancer team MDTs [40]. Factors impacting decisional quality included cancer management changes by individual physicians (2–52%), failure to reach a decision at MDT discussion (27–52%), failure of implementation (1–16%), limited engagement of nursing personnel, failure of consideration of patient preferences, time pressure, excessive caseload, low attendance, poor teamwork and lack of leadership leading to lack of information and deterioration of decision-making.
A 2009 survey of multidisciplinary team function within the UK National Health Service identified team composition, infrastructure, meeting organisation and logistics, patient-centred clinical decision-making and team governance as key domains of effective MDM function [41]. Despite this address to organisational process, a 2020 audit of lung cancer MDT function provided 10 challenging stage III cases with identical information to 11 different MDTs and identified substantial functional differences in outcomes in terms of agreement in TNM staging and treatment recommendations [42]. Systematic review of MDM quality-assessment tools reports variable coverage of the key domains of MDM function, finding little evidence of engagement for quality improvement [43].
While MDM has been broadly shown to increase the delivery of desirable guideline-recommended treatments to lung cancer patients, there are a number of other potential benefits, outside formal guidelines, which are not well described in the literature. Such benefits may include increased management efficiency, optimised patient–healthcare provider interactions, reduced consultation visits and duration, reduced inappropriate treatment (such as substitution of stereotactic ablative radiotherapy for resection in marginal operable cases), treatment plan optimisation, enhanced precision medicine outcomes, enhanced treatment plan communication, improved quality of life, improvements in patient and provider perceived healthcare value in addition to potential health economic benefits.
The benefits of MDM presentation are relatively well described in lung cancer (NSCLC and SCLC); however, there is scope to utilise MDM function to provide similar benefit in other thoracic cancers including thymoma, carcinoid tumours, neuroendocrine tumours and mesothelioma.
Study clinical heterogeneity
There was variation in included study participants reporting early stage I–II cohorts [34], locally advanced stage III [25], all-stage patients and inoperable lung cancer [23, 24]. Study reports included single institution reports [23, 29, 32], hospital networks [7, 20, 22, 27, 34], Veterans Affairs medical centres [26], state-based cancer registries [5, 33] and large national population registries [28]. MDM exposure included presentation to scheduled formal MDM [5, 7, 18–21, 23, 25, 26, 29, 30, 32, 33], and multidisciplinary care programmes [22, 24, 27, 28, 31], although timing of MDM presentation was not routinely specified. Study designs included evaluation before and after the implementation of multidisciplinary case meetings [22, 29, 32], contemporaneous cohorts retrospectively identified as MDM presented or nonpresented cohorts [5, 7, 17, 19, 21, 27], using landmark analysis [31] and propensity score matching of MDM presented/nonpresented [31]. Heterogeneity between observational studies is expected. We therefore planned to explore differences between studies through subgroup analyses by study design, by participants or by intervention type, which did not fully explain heterogeneity.
Strengths and limitations
This is the first substantive meta-analysis of management, treatment and survival outcomes in large populations of NSCLC patients treated in multiple national jurisdictions. While most studies described similar approaches in participation, conduct and timing of MDM meetings, it is likely that there was a degree of variation between process and outcomes of these meetings at participating centres, representing potential unwarranted variation in the intervention. The timing of exposure to the MDM process is not clearly reported and these factors may impact effective utilisation of neo-adjuvant and adjuvant therapies and become increasingly important in the modern era of immunotherapy and targeted therapies in early-stage disease [44].
Implications for practice, policy and future research
There remains a paucity of data available to describe MDM process, consensus strategies, evidence utilisation and adherence and audit outcomes in the implementation of MDM practice [45]. The literature evidence in multidisciplinary care function remains challenged by a wealth of observational study data and a paucity of prospective randomised trial evidence. There remains a lack of clarity in definitions of multidisciplinary care and process, and an absence of multidisciplinary care implementation strategies [46].
Points for clinical practice
• MDM presentation is a relatively low-cost intervention that improves histological confirmation, clinical staging, receipt of treatment and survival.
Questions for future research
• The determinants of MDM presentation are not well known.
• MDM presentation may have an important impact on cancer burden, improve equity and reduce unwarranted clinical variation in care.
Conclusion
The literature provides extensive evidence of benefit of MDM presentation, and yet substantial rates of nonpresentation exist with a lack of clarity of reasons for nonpresentation. Equity of access remains a significant concern, with clear evidence that certain populations may be underpresented, including the aged, those with stage IV disease, those living in nonmetropolitan centres and those treated in private healthcare systems.
Supplementary material
Supplementary Material
Please note: supplementary material is not edited by the Editorial Office, and is uploaded as it has been supplied by the author.
Supplementary table S1 ERR-0157-2023.table_S1
Supplementary analyses: outcomes by subgroup ERR-0157-2023.supplementary_analyses
Footnotes
Provenance: Submitted article, peer reviewed.
Conflict of interest: All authors have nothing to disclose.
- Received July 31, 2023.
- Accepted February 23, 2024.
- Copyright ©The authors 2024
This version is distributed under the terms of the Creative Commons Attribution Non-Commercial Licence 4.0. For commercial reproduction rights and permissions contact permissions{at}ersnet.org
References
