Baseline staging tests after a new diagnosis of breast cancer: further evidence of their limited indications

F. Puglisi1,2,*, A. Follador2, A. M. Minisini1,2, G. G. Cardellino1,2, S. Russo1,2, C. Andreetta1,2, S. Di Terlizzi2 and A. Piga2

1 Breast Unit and 2 Clinical Oncology, University of Udine, Udine, Italy

* Correspondence to: Dr F. Puglisi, Clinica di Oncologia Medica, Policlinico Universitario, Piazzale SM Misericordia, 33 100 Udine, Italy. Tel: +39-0432-559304; Fax: +39-0432-559305; Email: fabio.puglisi{at}med.uniud.it


    Abstract
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 References
 
Background: Bone scanning (BS), liver ultrasonography (LUS) and chest radiography (CXR) are commonly used in patients with newly diagnosed breast cancer as part of baseline staging. However, in the absence of symptomatic disease, the usefulness of this routine diagnostic work-up is not evidence-based.

Methods: We selected the study sample from 516 consecutive patients with newly diagnosed invasive breast cancer. For each diagnostic test (BS, LUS, CXR), we analyzed the prevalence defined as the number of patients with diagnosis of metastatic disease after an imaging technique divided by the total number of patients tested. In addition, sensitivity and specificity were calculated. Initial suspicion was confirmed by other independent tests (bone X-ray, computerized tomography scan, magnetic resonance imaging) in order to identify ‘true’ positive diagnoses.

Results: At baseline, BS was carried out in 412 patients, LUS in 412 patients and CXR in 428 patients. Thirty-three patients were correctly diagnosed by the initial staging investigations as having metastatic disease (true positive cases). BS detected skeletal metastases in 6.31% of patients, LUS detected liver metastases in 0.72% of patients and CXR detected lung metastases in 0.93% of patients. Before imaging tests, all patients with either LUS or CXR evidence of metastases were previously classified as having stage III disease. On the other hand, only 26.9% of bone metastases were detected in patients with stage III. Accordingly, the detection rate in stage III patients was 14%, 5.6% and 7.2%, respectively for BS, LUS and CXR.

Conclusions: These findings indicate that a complete diagnostic work-up to detect metastases is unnecessary in the majority of patients with newly diagnosed breast cancer, whereas it may be indicated for specific patient categories such as those with stage III disease.

Key words: baseline staging, bone scan, breast cancer, chest radiography, liver ultrasonography


    Introduction
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 References
 
An accurate determination of the stage is considered of crucial importance for the process of clinical decision making in patients with cancer, since knowledge of the extent of disease is strictly related to the patient's prognosis and has a strong influence on the therapeutic choices.

In patients with breast cancer, correctly staging the disease is important in guiding patients and physicians towards the most appropriate treatment option, namely to decide in favor of loco-regional versus systemic therapy [1Go, 2Go].

Traditionally, and according to the most common sites of metastases, non-invasive radiological tests routinely employed in the staging work-up for breast cancer have included bone scanning (BS), liver ultrasonography (LUS) and chest radiography (CXR). However, the yield of these tests has gradually scaled down after several studies reporting their inappropriateness especially in determinate subgroups of patients, such as those with small tumors and absent or minimal involvement of the axillary nodes [3Go, 4Go]. In fact, the use of these radiological tests in all breast cancer patients is unnecessarily expensive and time consuming. In addition, false positive results of the tests that require additional confirmatory examinations [computed tomography (CT), magnetic resonance imaging (MRI), positron emission tomography] may cause anxiety in patients and increase costs.

Nonetheless, many oncologists and breast surgeons probably continue to routinely order these tests as part of the baseline work-up. In addition, although with different implications and significance, the staging process by BS, LUS and CXR is always requested before enrolling a patient in a clinical trial. Furthermore, no official guidelines on this topic have been set out to date by the leading professional societies, such as the American Society of Clinical Oncology (ASCO) or the European Society of Medical Oncology (ESMO). Instead, ASCO clinical guidelines for breast cancer surveillance do not recommend serial imaging tests following primary intervention [5Go].

In order to add another piece of evidence on this topic, we reviewed findings about staging procedures from our database of breast cancer patients. In addition, unlike previous studies, we considered in the analysis some preoperative variables such as radiological size of the tumor, histological grade and hormonal receptor status on core biopsy, and the preoperative value of tumor marker CA 15.3. These variables together with others analyzed after surgery (pT, pN, tumor size) were assessed to identify specific subgroups of patients with a higher risk of metastatic disease (higher pre-test probability).


    Patients and methods
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 References
 
We selected the study sample from 516 patients with newly diagnosed invasive breast cancer who were consecutively referred to our institution (Clinical Oncology, University Hospital of Udine, Italy) between February 1999 and August 2003. The medical oncology records were reviewed to obtain data about patients and tumor characteristics (histology, stage, tumor grade on core biopsy, lymph node involvement, hormonal receptor status, radiological and pathologic tumor size, preoperative value of CA 15.3) staging tests (BS, LUS, CXR) and confirmatory investigations (bone X-ray, CT scan, MRI). LUS were carried out cases without contrast. For the aim of the study, the current version of the American Joint Committee on Cancer (AJCC) tumor–node–metastasis staging system was adopted [6Go].

For each staging procedure, we analyzed the ‘prevalence’ defined as the number of patients with diagnosis of metastatic disease after an imaging technique divided by the total number of patients tested. In addition, sensitivity (true positive results/true positive results + false negative results), specificity (true negative results/true negative results + false positive results) and positive predictive value (PPV: true positive results/true positive results + false positive results) were calculated. For the analysis, results categorized as suspect by the staging tests were computed as positive (higher sensitivity and lower specificity are therefore expected). Initial suspicion was confirmed by other independent tests as appropriate in order to identify ‘true’ positive diagnoses. In particular, bone scan abnormalities were evaluated by bone X-ray and/or CT scan and/or MRI. Abnormal LUS results were evaluated by CT scan and/or MRI. CXR abnormalities were studied by CT scan. Cases in which the diagnosis was not conclusive after further investigations were monitored for at least 6 months.

Associations between categorical variables were evaluated using the chi-square test or Fisher's exact test as appropriate. The nonparametric Mann–Whitney rank sum test was used for comparison of groups with ordinal data.


    Results
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 References
 
Among the 516 cases available for review, a BS was documented in 412 (80%), a LUS in 412 (80%) and a CXR in 428 (83%). The missing data were mainly due to the retrospective nature of the analysis: in the majority of cases (60%), the patients underwent surgery in our hospital but they were lost at follow-up; in 20% of patients, clinical signs of potential metastases were present, so they were deliberately excluded from the analysis; in the remaining 20%, only information about negative baseline tests was available, but medical reports were not available. Detailed information about stage was present in 462 (89.5%) of the records. Tumor characteristics are summarized in Table 1.


View this table:
[in this window]
[in a new window]
 
Table 1. Tumor characteristics

 
After the initial staging investigations, 33 patients were diagnosed as having metastatic disease. BS detected skeletal metastases in 26 (6.31%) patients, LUS detected liver metastases in 3 (0.72%) patients and CXR detected lung metastases in 4 (0.93%) patients. Data summarizing results with the three imaging techniques are listed in Table 2.


View this table:
[in this window]
[in a new window]
 
Table 2. Diagnostic accuracy in the diagnosis of metastatic disease by bone scanning, liver ultrasonography and chest radiography

 
Sensitivity, specificity and PPV of BS were 100%, 93.5% and 50.9%, respectively. Sensitivity, specificity and PPV of LUS were 100%, 93.6% and 10.3%, respectively. Sensitivity, specificity and PPV of CXR were 100%, 96.9% and 23.5%, respectively. Obviously, 100% sensitivity for each diagnostic test depends on the adopted criteria and it cannot be used by any means as a measure of accuracy. All patients with either LUS or CXR indication of metastases were previously classified as having stage III disease. On the other hand, only 26.9% of bone metastases were detected in patients with stage III disease. Accordingly, the prevalence in stage III patients was 14%, 5.6% and 7.2%, respectively for BS, LU and CXR (Table 3).


View this table:
[in this window]
[in a new window]
 
Table 3. Prevalence of metastatic disease by bone scanning, liver ultrasonography and chest radiography according to the pre-imaging stage

 
Among tumor characteristics, a statistically significant association was found between the number of involved axillary nodes and liver metastases detected by LUS (P=0.02). In addition, in two cases of liver metastases, a grade 3 tumor was observed on core biopsy (P=0.05). A marginally significant association was also found between high values of CA 15.3 and bone metastases (P=0.05).


    Discussion
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 References
 
At present, there is no clear evidence supporting the use of the traditionally employed imaging procedures (i.e. BS, LUS and CXR) to carry out baseline staging of breast cancer patients. On the contrary, several studies reported a limited value of breast cancer baseline staging, suggesting that complete diagnostic work-up should be limited to patients with higher pre-test probability of distant metastases [3Go, 4Go, 7Go–11Go]. In particular, it has been shown that the prevalence of detectable metastatic disease in the population of patients at the first diagnosis of breast cancer is exceedingly low and increases from stage I to stage III [3Go, 4Go]. In a systematic review, Myers et al. reported percentages of patients with positive BS at baseline varying across studies from 0.9% to 40% (all stages) with the lowest prevalence observed in stage I patients [0.5%, 95% confidence interval (CI) 0.1–0.9] and the highest in stage III patients (8.3%, 95% CI 6.7–9.9) [4Go]. Studies evaluating the role of LUS in detecting liver metastases showed an even lower prevalence than that observed in the studies of BS: 0% for stage I, 0.4% for stage II (95% CI 0–0.8), 2% for stage III (95% CI 0.4–3.6) [4Go]. The only two studies on CXR examined in the review reported similar results (0.1%, 95% CI 0–0.3 for stage I; 0.2%, 95% CI 0–0.4 for stage II; 1.7%, 95% CI 0.8–2.6 for stage III) to that observed with LUS.

In addition, Chen et al., in a large study (n=1493) designed to identify patients with early breast cancer who were upstaged to stage IV by a CXR, demonstrated a pulmonary metastasis prevalence of 0.099% in asymptomatic patients [7Go].

In a more recent paper, Ravaioli et al. reviewed data about staging procedures on 1218 breast cancer patients [3Go]. An overall prevalence of 1.46% was reported when results of BS, LUS and CXR were analyzed together in patients with low risk tumor characteristics (pT1–T3 and a number of positive nodes ≤3). On the other hand, in patients belonging to the high risk group (pT4 or a number of positive nodes >3 or pN2), the prevalence was 10.68%.

The authors concluded that new recommendations for the staging of breast cancer should be formulated, suggesting the use of BS, LUS and CXR only in patients with an a priori high risk of metastases.

Our study compares well with the data in the literature: patients with AJCC stage I and II breast cancer should not be routinely staged by LUS and CXR at baseline. In fact, the prevalence of liver and lung metastases detected by imaging techniques is equal to zero in the presence of early stage disease and dramatically increases in patients with stage III tumors (Table 3). Interestingly, in our series, a few cases of skeletal metastases were still diagnosed in stage I and II (prevalence of 5.08% and 5.55%, respectively), but no predictors were found to discriminate, better than stage, cases with higher risk of bone metastases.

The aim of our study was also to evaluate whether some preoperative variables such as radiological tumor size, tumor marker Ca 15.3, tumor grade on core biopsy or hormonal receptor status could predict the chance of detecting distant metastases. The availability of such earlier predictors should be very useful to avoid unnecessary surgery in patients with imaging confirmed metastatic disease. We found only a marginal association between higher values of Ca 15.3 and bone metastases and between higher tumor grade on core biopsy and liver metastases. These findings, however, must be interpreted with caution due to the limited sample size.

In conclusion, our study strengthens the recommendation to limit baseline staging tests, especially LUS and CXR, to patients at higher risk (stage III) of distant metastases. An indiscriminate use of diagnostic procedures in all patients with newly diagnosed breast cancer has a high chance of generating unjustified anxiety in patients and of increasing costs.

Finally, the unproven effectiveness of these tests in routine follow-up casts doubt upon their usefulness as a baseline reference [5Go].


    Acknowledgements
 
We wish to thank Dr Sabrina Guerra for her assistance in reviewing the manuscript. The results of this study were presented in part (poster session) at the 40th Annual Meeting of the American Society of Clinical Oncology, June 5–8, 2004, New Orleans, LA.

Received for publication July 12, 2004. Revision received October 21, 2004. Accepted for publication October 22, 2004.


    References
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 References
 
1. Delarue NC, Peters V, Anderson WS, Starr J. A re-evaluation of the place of major extirpation in the management of patients with metastatic mammary carcinoma. Can Med Assoc J 1967; 96: 637–651.[Medline]

2. Bedwinek J, Rao DV, Perez C et al. Stage III and localized stage IV breast cancer: irradiation alone vs irradiation plus surgery. Int J Radiat Oncol Biol Phys 1982; 8: 31–36.[ISI][Medline]

3. Ravaioli A, Pasini G, Polselli A et al. Staging of breast cancer: new recommended standard procedure. Br Cancer Res Treat 2002; 72: 53–60.[CrossRef][ISI][Medline]

4. Myers RE, Johnston M, Pritchard K et al. Baseline staging tests in primary breast cancer: a practice guideline. Can Med Assoc J 2001; 164: 1439–1444.[Abstract/Free Full Text]

5. Smith TJ, Davidson NE, Schapira DV et al. American Society of Clinical Oncology 1998 Update of Recommended Breast Cancer Surveillance Guidelines. J Clin Oncol 1999; 17: 1080–1082.[Abstract/Free Full Text]

6. Singletary SE, Allred C, Ashley P et al. Staging system for breast cancer: revisions for the 6th edition of the AJCC Cancer Staging Manual. Surg Clin North Am 2003; 83: 803–819.[ISI][Medline]

7. Chen A, Carlson GA, Coughlin BF et al. Routine chest roentgenography is unnecessary in the work-up of stage I and II breast cancer. J Clin Oncol 2000; 18: 3503–3506.[Abstract/Free Full Text]

8. Dillman RO, Chico S. Radiologic tests after a new diagnosis of breast cancer. Eff Clin Pract 2000; 3: 1–6.[Medline]

9. Samant R, Ganguly P. Staging investigations in patients with breast cancer: the role of bone scans and liver imaging. Arch Surg 1999; 134: 551–553.[Abstract/Free Full Text]

10. Gerber B, Seitz E, Muller H et al. Perioperative screening for metastatic disease is not indicated in patients with primary breast cancer and no clinical signs of tumor spread. Breast Cancer Res Treat 2003; 82: 29–37.[CrossRef][ISI][Medline]

11. Schneider C, Fehr MK, Steiner RA et al. Frequency and distribution pattern of distant metastases in breast cancer patients at the time of primary presentation. Arch Gynecol Obstet 2003; 269: 9–12.[CrossRef][Medline]





This Article
Abstract
Full Text (PDF)
E-letters: Submit a response
Alert me when this article is cited
Alert me when E-letters are posted
Alert me if a correction is posted
Services
Email this article to a friend
Similar articles in this journal
Similar articles in PubMed
Alert me to new issues of the journal
Add to My Personal Archive
Download to citation manager
Disclaimer
Request Permissions
Google Scholar
Articles by Puglisi, F.
Articles by Piga, A.
PubMed
PubMed Citation
Articles by Puglisi, F.
Articles by Piga, A.