The Second International Conference on Screening for Lung Cancer
Location: Weill Medical College of Cornell University
1300 York Avenue, New York NY 10021
Friday, February 25, 2000 to Sunday, February 27, 2000.
Summaries of the Workshops
Conference Missions: To seek consensus on:
Variants of the Early Lung Cancer Action Project (ELCAP) regimen of screening-diagnosis deserving consideration in updated optimization of the regimen.
Variants of the ELCAP intervention (lobectomy, singularly) deserving consideration in update optimization of the regimen.
Main principles of research aiming to provide the scientific knowledge-base for optimization of regimens of screening for lung cancer and quantification of their effectiveness - ultimately in preventing deaths from lung cancer.
Workshop I: Pooling of data
Chair: Marek Kimmel
Members:
Nathaniel Berlin, Daniella Braz, David Burns, John K. Field, Janet Husband, Deborah Marshall, Eugenio Paci, Thomas Scherer, Tomotaka Sobue, Shusuke Sone, Steven Springmeyer, Denise Warzel,
Background:
Central objects of the ELCAP research:
Diagnostic mission of the ELCAP is to quantify the extent to which early diagnosis of lung cancer is achieved by a defined regimen of screening-diagnosis (screening test together with the definitive diagnostics prompted by positive results of the test). This is given by the attained diagnostic distribution, i.e., distribution of cases diagnosed under screening (whether screen- or symptom-prompted) by stage -- principally a matter of frequencies of stage I cases subclassified by size. Principal interest in repeat screenings.
Interventive mission of the ELCAP is to quantify the frequencies of curability of diagnosed lung cancer, across the defined regimens of screening-diagnosis. This is principally given by the 'curability function', that is the rate of curability as a function of stage/size of the tumor, notably size in stage I cancer. A critical distinction is to be made between screen- and symptom-diagnosed cases, and with principal interest. The principal interest in repeat screening. Problem: In the ELCAP, paucity of experience with repeat screening. Same applies to similar projects elsewhere.
Solution:
Pooled analysis of separate projects, preferred over meta-analysis of separately reported results.
Mission:
To develop draft statements of consensus on the following:
1. Is the pooling desirable? Reasons.
2. Insofar as it is desirable,
a) what principles bear on poolability?
b) what other guidelines should be appreciated in the pooling project?
Workshop I recommendations:
We want to test whether the single-arm study is a valid approach to evaluation of screening for early detection of lung cancer, which can answer the important questions faster than a randomized clinical trial. Recommendations reflect the consensus on pooling reached in this context.
Purpose of pooling data from different studies. One principal goal of pooling is to increase the sample size available for estimation of the yield of incidence screens, tumor growth rates and clinical outcomes of those screened. Another purpose is to create a database sufficient to answer present and future questions. Examples of such questions are those concerning the natural history, detection and treatment of lung cancer. Pooled data can be used to build a mathematical model of disease, screening and treatment, which will generate projections of the efficacy of screening.
Organizational issues. An inter-institutional board should exist to issue recommendations on standardization of data and analysis standards, as well as to preserve equilibrium between proprietary rights of researchers and the ability to use pooled data. The board will take responsibility for the maintenance, confidentiality and data access authorization. The board could function in a wider framework and be linked to international institutions.
Nature and compatibility of data pooled. The basic list includes data on screening protocol, detection modality, digital CT images, pathology, staging, treatment, follow-up and epidemiological data. Data from different institutions will reflect differences in the intrinsic nature of disease, sensitivity of screening, pathologic evaluation, therapeutic intervention, costs of medical care and other issues. Standardization before pooling is a necessity. Preservation of tumor specimens for future molecular analyses should be recommended. Consistency with NCI Common Data Element standards is desirable. The database should include the collection of tissue and sputum samples to allow us to test a range of molecular-pathological biomarkers, which identify high risk individuals.
Workshop II: Research on diagnostic components in screening research
Background:
'Screening' in the sense of a diagnostic regimen - screening test together with the definitive diagnostics prompted by positive results of the test, periodically repeated - involves a number of components; and for practice, the concern is to formulate the regimen - the set of components - with a view to optimization of cost-effectiveness in the pursuit of 'early' diagnosis. Different screening tests lead to different diagnostic distributions (by stage and size), while different definitive diagnostics lead to the same diagnostic distribution at different costs and burdens on the screenee (risks included). Practice-oriented research on screening thus involves study of the relative merits of potentially optimal options in the components of screening diagnosis.
Mission 1:
In CT screening for lung cancer, with focus on repeat screenings, which aspects of the diagnostic regimen (from the screening test all the way to definitive diagnosis) are in most urgent need of research-based development? Assessment of growth in smallest nodules? Indications for biopsy? Understanding the 'ground-glass opacities' vis-a-vis truly suspicious nodules?
Mission 2:
In this diagnostic research, is there a role for RCTs? Reasons?
Chair: David Yankelevitz, David Naidich
Members:
Robert Clark, Peggy McCarthy, Stefan Diederich, Stanley Fox, Salvatore Giunta, Ali Farooqi, Ryutaro Kakinuma, Mashadiro Kaneko, Stephen Lam, Eugenio Paci, Branko Palcic, Thomas Petty, Anthony Reeves, Dorith Shaham, James P. Smith, Al Tierstein, Irena Tocino, Michael Unger, Steven Weiss, Bob Webster, Javier Zulueta and other participants of the conference
Recommendations of Workshops II:
It was agreed that the single-arm type of protocol for the diagnostic workup is an evolving process responsive to evolving technology. It was recognized that beyond CT screening, cytological screening using sputum would be useful in both a complementary and comparative role. Lung cancer screening requires a multi-disciplinary approach with a view towards developing a consensus among key societies.
Figure 1 illustrates the flexibility of the single-arm screening-diagnostic mission. Critical to the validity of the single-arm protocol is that a specified low-dose CT screening protocol is followed, that the final diagnosis of each nodule becomes known (either by follow-up CT demonstrating no growth or pathologic diagnosis). The entry criteria should be specified but can be different at each institution. Similarly, the diagnostic workup can be different but the recommendations given to the physicians should be spelled out. All subjects diagnosed with malignancy should be followed for long-term outcome. Subjects should be invited back for annual repeat screening. Critical information which needs to be documented for purposes of evaluation is considered by the other workshops.
Figure 1. A schematic diagram of the single-arm (ELCAP) protocol
Workshop III: Research in Intervention Issues in Screening Research
Chair: Robert Ginsberg and James Mulshine
Members:
Nasser Altorki, Hisao Asamura, Christain Brambilla, Tom Daniel, Andrew Dannenberg, Roy Herbst, Fred Grannis, Harvey Pass (Rapporteur), Valerie Rusch, Junji Yoshida
Background:
The purpose of the pursuit of early diagnosis by screening for lung cancer is to provide for early intervention. The underlying premise is that early intervention is more commonly curative in asymptomatic screen-identified cancers than would be expected with symptom-prompted diagnosed lung cancer. CT screening brings the intervention issues to the domain of the very small lesions. In this domain, particularly when the lesion size is less than one centimeter, new issues arise:
1. When is resection called for?
2. Given that resection is required for malignancy, is it to be lobectomy, or might a lesser resection be adequate?
3. Given the high risk for a subsequent, new primary lung cancer, is chemoprevention worthwhile? And if so, what form(s) might it now take?
4. Is there a subdomain such that, despite pathologic diagnosis of malignancy, experimental study of the relative merits of immediate resection and no immediate resection is scientifically warranted and ethically justifiable? If so, what? Either way, what is the rationale?
5. Are there non-surgical methods worthy of study in treating subcentimeter lesions?
Mission:
Development of potential consensus statement(s) on any, or all, of the questions above.
Workshop III recommendations:
In view of the charge to develop prospective guidelines for management of patients enrolled in the International Collaboration to Screen for Lung Cancer, this committee focused exclusively on issues concerning peripheral lesions as these are the majority of lesions detected by spiral CT. Centrally located lesions should be evaluated using additional techniques including fiberoptic bronchoscopy and sputum cytology. Discussion of the management of these central lesions was not part of the workshop's scope, but should be pursued further.
The group was very aware that the broader use of spiral CT would result in an increased number of thoracic procedures involving very small primary lesions. One likely consequence includes the possibility of an increased rate of surgical procedures for benign disease with attendant cost and complication rate. This should be closely monitored
Another concern is the need to conserve lung tissue in individuals who might require serial surgical procedures for metachronous primary lung cancers. Since critical clinical information in this new area is only beginning to emerge, health care providers need to proceed thoughtfully to ensure that the care for each individual patient is optimized.
It is strongly recommended that screening of high-risk individuals be done in the setting of a center committed to providing the full range of relevant interdisciplinary support (either directly or through collaboration) for the management of each screen-identified case. This includes relevant specialists with interest in lung cancer. Activities should be coordinated using interdisciplinary management meetings. The I-ELCAP group should develop data sheets to capture critical information about the management of these early lesions in regard to both the diagnostic work-up and definitive treatment including complications.
The entire screened group should be closely followed. All patients following a first lung cancer treatment need to be continually monitored for subsequent lung cancers. Although there is no current evidence to recommend a particular follow-up policy using imaging modalities after resection, the committee suggests that cases identified in a CT screening program should be maintained in a CT surveillance program. This recommendation is prompted by a 1-3% per year cumulative risk of new primary lung cancer.
Evaluation of molecular characteristics of resected malignancies should be included whenever possible.
Chemoprevention research should be strongly encouraged and provisions to pool the resulting data into the I-ELCAP database should be implemented.
The standard of care for pathologically confirmed lung cancer universally usually includes definitive treatment. If a patient refuses or is unable to receive definitive therapy, that patient should be followed for ultimate outcome.
1. Intervention for lesions of size 5mm or less
In the prevalence screening setting, the group supports the current ELCAP recommendation for observation of lesions 5mm or less with a high resolution CT evaluation at 3-month intervals without immediate surgery or further investigation.
The criterion for diagnostic intervention is interval growth on follow-up CT. Prior to definitive treatment it is desirable to confirm the diagnosis of cancer. The diagnostic procedure to confirm malignancy should be the least invasive that is within the capabilities of the local care providers (although expertise with needle biopsy should be encouraged, it is recognized that this is not universally available). The alternative procedures in the absence of expertise in image-directed needle biopsy include bronchoscopic maneorves, thoracoscopy and/or thoracotomy.
Once diagnosed, the current standard of care is lobectomy and lymph node sampling or dissection. Wedge resection should be discouraged except in special circumstances. In the case of a patient who is managed by a wedge biopsy that is later found to be invasive cancer, the treatment of choice is a completion lobectomy with lymph node sampling provided that the pulmonary status is adequate. Other possible options in this situation include i) mediastinoscopy to define nodal status with watchful waiting in the absence of a positive mediastinoscopy, or ii) post operative radiation therapy with fields determined by the results of the mediastinoscopy. Surveillance in this situation may need to be more frequent than usual, and may include serial CT scans, the frequency of which should be determined by the individual institution to look for both new cancers and local recurrence. This is an important area for clinical research. Wedge resection with operative mediastinal staging may be acceptable for the pulmonary-compromised individual. Further local control measures such as intraoperative interstititial brachytherapy to extend the surgical margin of the wedge resection are reasonable. In addition primary radiation therapy using intraoperative interstititial brachytherapy or post-operative focused exxternal beam radiation, conformal external beam radiation using intensity modulated therapy or stereotactic radiosurgery with bodyframe can be employed. Conventional radiation therapy with regional boost is not recommended.
In a research setting other procedures such as segmentectomy or wide wedge resection could be investigated. The group encouraged such research approaches of less invasive procedures in the setting of very small-volume primary lung cancer or bronchioloalveolar carcinoma because there is not a large experience base. Therefore pooling of clinical research experience with various management approaches is essential.
2. Intervention for lesions of size 6-10 mm
In the prevalence setting, a growing lesion or a lesion deemed suspicious for malignancy by imaging criteria, a diagnostic procedure to rule out malignancy should be obtained. New lesions detected on an incidence screen should be diagnostically investigated unless there is unequivocal evidence for benignity. The diagnostic procedure to confirm malignancy should be the least invasive procedure that is within the capabilities of the local care providers (although expertise with needle biopsy should be encouraged, it is recognized that this is not universally available). In the ELCAP experience, confirmation of malignant disease could be obtained by fine needle aspiration biopsy in over 80% of cases prior to surgical resection. Therefore, the committee discourages the use of diagnostic surgical procedures unless absolutely necessary. This group encourages the development of less invasive approaches to the diagnostic evaluation 6-10 mm CT-detected lesions. The alternative procedures in the absence of expertise in image-directed needle biopsy are thoracoscopy and/or thorocotomy. The current standard of care is lobectomy with lymph node sampling for these lesions, but lesser resection may be an option as outlined in section 1.
3. Interventions for lesions greater than 1cm
In the setting of a greater than 1 cm lesion, standard cancer evaluation and management is recommended.
Workshop IV: Quality assurance in screening research
Background:
Research contributing to the knowledge-base of decisions about screening for lung cancer is to provide evidence of adequate quality for this purpose. To this end, it is to address questions that truly are relevant to optimization of the regimen(s) of screening in the foreseeable future, and/or to quantitate its (their) effectiveness in the ultimate sense of prevention of deaths from lung cancer and/or some component issue(s) bearing on this; and it is to be of adequate validity. (Precision is a matter of quantity, not quality, of evidence).
Mission 1:
Development of potential consensus statement(s) on the principal requirements for adequate quality in an ELCAP-type study, and on quality assurance in respect to these.
Mission 2:
Development of potential consensus statement(s) on the principal requirements for adequate quality in an RCT approach.
Chair: David Yankelevitz, David Naidich
Members:
Robert Clark, Peggy McCarthy, Stefan Diederich, Stanley Fox, Salvatore Giunta, Ali Farooqi, Ryutaro Kakinuma, Mashadiro Kaneko, Stephen Lam, Eugenio Paci, Branko Palcic, Thomas Petty, Anthony Reeves, Dorith Shaham, James P. Smith, Al Tierstein, Irena Tocino, Michael Unger, Steven Weiss, Bob Webster, Javier Zulueta and other participants of the conference
Recommendations of Workshops IV:
A teaching file, for future physician education should be established with a lexicon of the standard nomenclature for nodule features. Interaction with the groups developing the common data elements being established by the NCI is desirable.
The quality assurance aspects required for multi-institutional screening evaluation using CT focused on the following broad topics:1) inclusion criteria, 2) CT screening technique parameters, and 3) further diagnostic workup.
1) Inclusion Criteria
The ELCAP approach allows for variability in the inclusion criteria. In fact, such variability is desirable as it provides for assessment of alternative approaches.
a. Inclusion criteria would be set by each institution with documentation of the age, smoking history, health status (functional impairment, symptoms, operability).
b. Cost information should be collected.
c. All subjects should have a physician of record so that the subject and physician can be notified about the findings by the radiologist.
d. Screening should be performed under an approved research protocol (e.g., informed consent including potential limitations of the technique).
2) Screening CT
The screening techniques used by seven institutions currently performing screening were reviewed. Although there was variability, general scanning parameters (for both helical and electron beam scanners) can be established for both initial screening studies as well as the diagnostic follow-up CTs.
a. Low-dose scanning (minimally defined as 40-80 mA at 2:1 pitch) should remain the standard screening test.
b. High-resolution images of nodules identified by the screening CT was considered optimal for further nodule characterization and growth determination.
c. Annual repeat CT screening was reasonable at this time.
d. Double readings of the screening CT was recommended with documentation of the individual readings and the final consensus reading.
3) Diagnostic Workup
It was agreed that the ELCAP approach allows for flexibility in the workup (e.g, CT surveillance, antibiotic therapy, lung biopsy including transthoracic needle biopsy, bronchoscopy, and thoracotomy). Variability among institutions as to the diagnostic workup is desirable as it provides for assessment of alternative approaches. It was recognized that for the validity of the ELCAP approach, the final diagnosis has to be firmly established and documented, either by resection or follow-up.
Measurement of the size and volume should be as precise as possible using computer methods. It was agreed that emphasis be placed on the development of 3D rendering, volume determination, and nodule detection and data should be stored for future testing of new techniques.
Workshop V: Quality Assurance of Pathology in Screening Research
Background:
In screening for lung cancer, pathologic diagnosis is, at present at least, the positive end result of the screening-diagnostic process; and by the same token, it is pivotal to decisions about the early interventions that the screening is intended to provide for. For this reason, quality of the pathology aspects of screening research deserves separate and special attention in relation to the general concerns in Workshop IV.
The general question here is: how should pathologic specimens be obtained, described/classified, and stored in a screening study in order that the processes represent potentially optimal future practice for one, and accord with the objectivity/reproducibility requirement of science for another - and provide for future reclassifications besides.
Mission:
Development of potential consensus statement(s) in the spirit of the Background statement above.
Chair: William Travis
Members:
Darryl Carter, Adi Gazdar, Fred Hirsch, June Koizumi, Masayuki Noguchi, Douglas Flieder (Rapporteur), P Snijders, Erik Thunnissen, Klaus Schreiber, Madeline Vazquez
Recommendations:
The advent of CT screening for early detection of lung cancer presents an ideal opportunity to address important questions about the pathology and biology of lung cancer. This is particularly true for small adenocarcinomas showing a non-invasive bronchioloalveolar pattern which have recently been shown to be potentially curable. Since there is little clinical data based on carefully characterized small carcinomas and putative precursor lesions, standardized protocols for collection and evaluation of these lesions along with review of all cases by an expert panel are strongly recommended. While this panel recognizes that screening-detected cancers require a more extensive cyto-pathologic evaluation than symptom-detected "routine" cancers, it is critical that a full pathologic evaluation of the tissue be conducted for study cases. It is fortunate that we now have the technology that allows for molecular and immunohistochemical studies from paraffin embedded tissue. Microdissection and microarray technology will permit studies to be done on precisely defined histologic changes.
Cytologists involved in CT-guided needle biopsies, i.e., fine needle aspirates (FNA) procedures, must endeavor to collect alcohol-fixed material as well as Diff-Quik, PAP smear and cell block preparations for future molecular studies. Wedge biopsy specimens should be fixed in formalin prior to staple removal, following staple removal and following thin (0.5 cm) sectioning. Shaking of the formalin-filled container and even the addition of a small amount of carbonated water is recommended to prevent specimen atelectasis. Resection specimens should be inflated through the main bronchus or bronchi and fixed for at least 24 hours prior to sectioning at 1.0 cm intervals. Careful mapping of the tumors to allow for tumor reconstruction and CT-scan 3-D correlation will preclude fresh tissue procurement. Ideally, extensive specimen sampling in conjunction with an inflated specimen CT scan could be performed; however, economic and time constraints will in most cases preclude such an undertaking without appropriate funding for the pathology support. Nevertheless, tumors and all other visible or palpable lesions should be completely examined histologically along with at least ten random sections of non-neoplastic lung and cartilage-bearing airways. All lymph nodes must be included. Extensive sampling with careful coding of the tissue blocks will allow for tumor pattern reconstruction and accurate size determination which must be correlated with CT size and 3-D shape parameters. In some cases this may result in over 50 tissue blocks. Extensive sampling of non-neoplastic lung will allow for re-evaluation of the current TNM staging system which considers a resected lung lobe harboring a microscopic satellite lesion in addition to the primary tumor to be a TNM T4 and thus stage IIIB cancer.
Cytology specimens should be evaluated for the presence or absence of malignancy with attention to small cell versus non-small cell type. Other acceptable diagnoses include infection, benign and atypical. Since a preoperative diagnosis of malignancy is required by most surgeons, atypical lesions should be reviewed by more than a single cytologist and perhaps studied by an expert panel in an attempt to make a definitive diagnosis.
Reports on resection specimens should include tumor type, size, cell type, degree of differentiation, nuclear grade, presence or absence of stromal invasion, presence or absence of pleural invasion, presence or absence of angiolymphatic invasion, presence or absence of satellite lesions (with complete descriptions of these lesions) and presence or absence of putative precursor lesions (atypical adenomatous hyperplasia, squamous metaplasia and/or dysplasia and pneumocyte proliferations or uncertain malignant potential). A data collection worksheet can simplify this task.
It must be noted that while the diagnosis of lung cancer is made with a high degree of accuracy by most cytologists and surgical pathologists, disagreements exist in some cases regarding, for example, tumor size, presence or absence of stromal invasion and the diagnosis of atypical adenomatous hyperplasia (AAH). In order to collect data on these controversial areas, all tumors should be measured grossly and microscopically and as mentioned above, the results compared with stated CT sizes. The acellular mucinous component of mucinous tumors and cystic component of partially cystic tumors should be included in size measurements. Tumor disruption by previous diagnostic procedures, i.e., FNA needle tract(s) should also be noted. Special studies including histochemical stains, i.e., VVG and Movat, as well as immunohistochemical stains, such as laminin and type IV collagen may facilitate the distinction between stromal invasion and secondary alveolar collapse. Subtle morphologic differences discern tobacco-related findings including respiratory bronchiolitis and reactive atypia from AAH. A more complete characterization of these lesions is only possible when non-neoplastic lung is extensively sampled.
It is this panel's recommendation that slides and representative tissue blocks be reviewed by an expert panel according to the 1999 WHO Histologic Classification and stored at the same central facility where CT scans are reviewed and stored. In cases with multiple lesions, the consensus diagnosis for each lesion should be recorded. Such will allow for standardized diagnoses and a tissue bank for envisioned morphometric, immunohistochemical and molecular characterization of the tumors.
<< Back to The Second International Conference on Screening for Lung Cancer General Information
Stay Up-to-Date!
Sign up below to join our email list and receive notice of site updates and other I-ELCAP news.
Contact Us
Have questions? Contact us by email, telephone or mail.
Get Screened
Do you need to reach one of our member sites? Search our site locator, or skim our full list of member sites.
Thank You
I-ELCAP would like thank Idologic, Inc., for its donation of hosting and support services.