- © 2000 by American Society of Clinical Oncology
Routine Chest Roentgenography Is Unnecessary in the Work-Up of Stage I and II Breast Cancer
- E. Alexandra Chen,
- Gregory A. Carlson,
- Bret F. Coughlin,
- William P. Reed Jr,
- Jane L. Garb and
- James L. Frank
- Address reprint requests to James L. Frank, MD, Department of Surgery, Baystate Medical Center, Springfield, MA, 01199; email james.frank{at}bhs.org
Abstract
PURPOSE: Clinical practice guidelines of many professional societies call for routine staging chest x-rays (SCXR) for all patients with invasive cancer. Given the estimated 157,000 patients annually for whom this recommendation pertains, this screening examination represents a considerable health care expenditure. If it were shown that SCXR rarely changed the management of low-risk subsets of this population, it might be possible to selectively omit this practice from the care of these patients with substantial resultant cost savings.
PATIENTS AND METHODS: All patients with clinical stage I and II breast cancer presenting to the Baystate Medical Center from 1989 through 1997 were identified through the Tumor Registry. Their hospital records were reviewed for clinical presentation and documentation of SCXR.
RESULTS: One thousand four hundred ninety-four patients were identified with clinical stage I and II disease. SCXR were available for review on 1,003 patients. Only one asymptomatic patient was upstaged to stage IV based on a SCXR. Two patients with primary lung tumors were also identified. These data demonstrate an asymptomatic pulmonary metastasis detection rate of 0.099% (95% confidence interval, 0.0% to 0.6%). The total charges of SCXR for this group approached $180,000.
CONCLUSION: These data demonstrate the low diagnostic yield and high cost of routine SCXR in the management of asymptomatic patients with clinical stage I and stage II breast cancer. Because other studies have shown that SCXR changes neither quality of life nor overall survival, SCXR should be limited to symptomatic patients in whom metastatic disease is suspected.
THE CARE OF cancer patients is becoming progressively more complex and costly. In response to increasing demands to control costs from government and private third-party payers and employers, health care providers are re-examining their current methods of patient care to optimize the utilization of resources. While addressing the need to pursue cost containment, oncologists must also fulfill their principal role as patient advocate in protecting the public from the untoward outcomes (eg, delay in diagnosis or inaccurate staging) that may follow strategies promoting profit over patient safety. Several studies, however, suggest that expensive staging and surveillance procedures can be safely eliminated.1-5 Bone scans, once considered a necessary part of the work-up of all patients with invasive breast cancer, have been virtually eliminated from the care of asymptomatic patients with small operable tumors.6-10 Staging chest x-rays (SCXR), however, continue to be recommended for all patients with invasive breast cancer even though pulmonary metastases are the initial site of metastasis in only 5% to 15% of patients11
We postulated that it might be possible to identify a low-risk subset of patients for whom omission of SCXR might be appropriate. The current study was undertaken to re-evaluate SCXR in patients with early (stage I and II) breast cancer, a subset of patients in whom the diagnostic yield and therapeutic benefit has been suggested to be minimal.
PATIENTS AND METHODS
All stage I and stage II breast cancer patients diagnosed between 1989 and 1997 were identified by the Baystate Medical Center Tumor Registry (group A). Inpatient and outpatient charts and hospital-based radiation therapy records were reviewed to identify documentation of an initial SCXR. Data retrieval from the multiple, geographically disparate outpatient settings (eg, private physician offices and other hospitals) where these patients underwent subsequent postoperative management was not pursued. Because this study involved patients entered onto the tumor registry over a 9-year period, the current version of the American Joint Committee on Cancer tumor-node-metastasis staging system during the year of entry into the Tumor Registry was used. A computerized search of tumor registry data for subsequent pulmonary recurrences of the stage I and II patients yielded only 14 patients, suggesting that recurrences might be underreported. Manual extraction of this data was not feasible, and so site of first recurrence could not be analyzed in this series.
To identify patients with early breast cancer who were upstaged to stage IV by an SCXR, all patients diagnosed with stage IV breast cancer were identified by the tumor registry (group B). Particular emphasis was placed on the identification of patients with small primary tumors and lymph nodes who lacked pulmonary symptoms and who had isolated pulmonary metastases detected by SCXR. The hospital’s inpatient and outpatient records of these patients were independently reviewed by two physicians (J.L.F. and G.A.C.), who demonstrated complete concordance in their review.
The total cost and charges for routine outpatient chest x-rays and chest computed tomographies (CT) were calculated based on the number of evaluations and cost/charge per study as prepared by Baystate Medical Center Decision Support Services. Hospital costs, which reflected resource consumption, were determined using a weighted simultaneous cost allocation method. Fiscal year 1999 cost and charge data were used.
RESULTS
Group A consisted of 1,493 patients. Of these, 1,352 (91%) had medical records available for review; 141 records could not be located. An SCXR was documented in 1,003 available cases (74%). Five patients had abnormal SCXR and underwent diagnostic chest CT to further evaluate the lesions detected on plain film. On the basis of the chest CT, two patients were determined to have primary lung cancers. Three patients were found to have benign entities such as granulomata.
Group B consisted of 82 patients with stage IV disease. Of these, 23 patients presented with simultaneous locoregionally (American Joint Committee on Cancer stage III) advanced breast tumors and concomitant metastatic disease; 58 patients presented with smaller (T1-2) primary tumors but had signs or symptoms suggestive of systemic metastases (Table 1). One asymptomatic patient was found to have bilateral pulmonary metastases initially detected by a preoperative SCXR. The patient was treated with primary chemotherapy after excisional biopsies revealed ductal carcinoma. Adding this patient, whose initial clinical examination demonstrated stage II disease, to group A, the total number of patients with clinical stage I and II breast cancer is 1,004. Using this number as a denominator, the asymptomatic pulmonary metastasis detection rate is 0.099% (95% confidence interval, 0.0% to 0.6%).
Clinical Presentation of Stage IV Patients, 1989–1997
At our institution, fiscal year 1999 costs of outpatient SCXR and chest CT (without intravenous contrast) are $62.92 and $118.27, respectively. Patient charges, including physician professional fees, are $181.01 and $696.70 for SCXR and chest CT, respectively. Using these 1999 rates, the total costs and charges for initial SCXR at this institution over the study period would have exceeded $63,000 and $185,000, respectively.
DISCUSSION
This study updates and confirms the findings of Ciatto et al12 who reported on 2,816 Italian patients diagnosed with stage I and II breast cancer between 1973 and 1985. In that series, SCXR identified only four patients with pulmonary metastases (detection rate 0.14%) The present study, conducted in an era of progressively smaller, mammographically detected tumors, provides a more contemporary perspective on the diagnostic yield of staging asymptomatic patients. Similar findings were also reported in two earlier, smaller series whose authors recommended that SCXR be reserved for symptomatic patients.13,14 Although most clinicians concur that diagnostic imaging is indicated to evaluate a patient’s symptoms, only a small proportion of SCXR are ordered for this purpose.14 Recommendations to omit routine chest imaging of asymptomatic patients become particularly compelling in the context of ample data demonstrating that intensive postoperative surveillance strategies to detect asymptomatic pulmonary metastases improve neither survival nor quality of life.15-19 Recognizing this, the American Society of Clinical Oncology clinical guidelines for breast cancer surveillance do not recommend serial CXR.20
Another argument for obtaining serial SCXR, that a patient’s emotional well-being benefits from a more intensive follow-up program, is not supported by the available data.18 There is also no cogent medical indication for routine preoperative chest x-rays for asymptomatic breast cancer patients undergoing general anesthesia. Although many anesthesiologists traditionally required this exam before administration of general anesthesia or conscious sedation, such a practice is no longer deemed necessary for otherwise healthy patients with no signs or symptoms of cardiopulmonary disease.21
SCXR detected two patients in this series with primary lung cancer. Depending on length of follow-up, second primary tumors of all sites are detected in 0.83% to 8% of breast cancer patients.22-25 Studies examining multiple primary cancers, however, are unable to establish a link between breast and lung tumors.26-28 Adequate data exist that lung cancer screening of the general public is ineffective.29 Thus, no argument can be made for obtaining SCXR to screen breast cancer patients for primary lung tumors.
Despite the complete lack of data supporting the use of this screening examination, SCXR are routinely ordered for all patients with invasive breast cancer, regardless of stage. Professional organizations, including the National Comprehensive Cancer Network, American College of Surgeons, and Society of Surgical Oncology recommend SCXR as part of the preoperative work-up.30-32 In many instances, such guidelines are based on the consensus opinions of their panel members, rather than relying on statistically valid, published data. Nonetheless, a recent survey revealed that most clinicians follow their professional society guidelines and order SCXR.33 Unless such recommendations are revised, it is unlikely that extant patterns of care will be altered. It is our opinion, based on the available data, that SCXR be ordered only for the following two groups of patients: those with pulmonary symptoms or those with clinical stage III or IV disease where the disease burden portends a higher diagnostic yield for the screening exam.11,12
The fiscal implications of these data for conserving health care resources are clear. A crude charge projection for the initial SCXR for 130,000 stage I and II patients would exceed $23 million annually. Based on our asymptomatic pulmonary metastasis detection rate, approximately 130 patients might be upstaged annually by an initial SCXR. Because these patients may eventually require therapy (surgery and/or radiotherapy) for locoregional control, the expense of SCXR cannot be justified on the basis of eliminating costly local therapies. Put another way, this represents an expenditure of at least $175,000 per patient for a diagnostic intervention that confers no survival benefit. We believe it is preferable to treat patients with disseminated metastases with systemic therapy first but question whether the cost of SCXR to evaluate all patients with stage I and II disease can still be justified.
Our detection rate of one in 1,004 patients (0.099%) may actually be a conservative one. Although 33% of the stage I and II patients in this series had no documentation of SCXR, it is likely that nearly all of these patients had this exam. During the study years, our hospital mandated that surgeons obtain a preoperative SCXR for all patients with invasive breast cancer. It was also considered a standard of care in the local medical oncology community to routinely order postoperative SCXR if one had not been performed preoperatively. That such SCXR were not documented in our inpatient records reflects the fact that many SCXR were obtained in other hospitals and outpatient clinics. Had an outside SCXR detected an asymptomatic pulmonary metastasis, it is highly likely that the patient would have been detected in our review of group B. Our detection rate of asymptomatic pulmonary metastases could thus be as low as 0.067%.
It is possible that our study missed patients with asymptomatic pulmonary metastases whose surgeon and medical oncologist both omitted SCXR from their work-up. To invalidate our principal findings, however, a major, undetected deviation from standard clinical practice would have to have occurred. Although we initially tried to identify such patients through a tumor registry search of site of first recurrence in group A, this search yielded only 14 patients, casting doubt on the accuracy of the tumor registry follow-up data. More accurate follow-up data might also have allowed us to identify patients whose initial SCXR missed subtle radiologic findings destined later to bloom as obvious pulmonary metastases. It could also be argued that, for this group, the initial SCXR was useful as a baseline study to document the subsequent change. As previously mentioned, this practice can not be justified on the basis of improving any patient outcome. Although we acknowledge the limitations imposed by the retrospective nature of the study’s design and limited medical record availability, a prospective study to evaluate SCXR seems no longer appropriate given the extant, compelling data of this and other series.
In summary, the routine SCXR has a low diagnostic yield and has never been shown to improve any outcome in the care of patients with clinical stage I and II breast cancer. There are no compelling nonstaging indications, such as for safe anesthesia, to recommend its use. The elimination of SCXR for asymptomatic patients with clinical stage I or II breast cancer will save valuable resources.
Acknowledgments
Supported by the Rays of Hope Foundation, Springfield, MA.
ACKNOWLEDGMENT
We thank Dr Wilson Mertens for reviewing the manuscript and Ms Roxy Grenier of the Baystate Medical Center Medical Records Department whose efforts made completion of the project possible.
Footnotes
-
Presented at the annual meeting of the Society of Surgical Oncology, Orlando, FL, March 14, 1999.
- Received January 18, 2000.
- Accepted June 6, 2000.
