Clinical significance of noncalcified lung nodules in patients with breast cancer

Transcription

1 Breast Cancer Res Treat (2016) 159: DOI /s CLINICAL TRIAL Clinical significance of noncalcified lung nodules in patients with breast cancer Feng Li 1 Samuel G. Armato 1 Maryellen L. Giger 1 Heber MacMahon 1 Received: 16 May 2016 / Accepted: 3 August 2016 / Published online: 9 August 2016 Springer Science+Business Media New York 2016 Abstract Patients with breast cancer are increasingly likely to have chest computed tomography (CT) performed. In many cases, small lung nodules will be detected, raising concern for metastases and causing considerable patient anxiety. The majority of these nodules, however, are benign, though the specific probability of malignancy is uncertain in any given case. Therefore, we analyzed the results of chest CT scans of a large number of patients with breast cancer, to determine characteristics and clinical significance of noncalcified lung nodules patients were investigated, and 4889 CT scans from 1325 patients were retrospectively reviewed. Among the 1325 patients, 812 (59 %) had at least one noncalcified lung nodule, of which 330 (41 %) had malignant nodules, 197 (24 %) had large (C10 mm) nodules, and 586 (72 %) had multiple nodules. Large nodules were more often malignant than benign (P \ 0.001). In patients with multiple large nodules, the rate of malignancy rate was 83 %, and most of these were metastases. In the case of very small (2 4 mm) nodules, the malignancy rates for solitary and multiple nodules were 8 and 20 %, respectively. Lung metastases were more likely with breast cancer cell grade 3 (22 %) than grade 1 2 (10 %) (P \ 0.001) and when patients were clinical stage 2 3 (14 %) than stage 0 1 (7.9 %) (P = 0.03). Lung metastases are highly likely in patients with multiple nodules greater than 10 mm. Higher cancer cell grades and clinical stage are also related to an Presented in part at the RSNA 2014 meeting. & Feng Li feng@uchicago.edu 1 Department of Radiology, The University of Chicago, 5841 S. Maryland Avenue, MC 2026, Chicago, IL 60637, USA increased likelihood of lung metastases. The great majority of small lung nodules in breast cancer patients are benign. Keywords Breast cancer Chest CT Lung nodule Malignant or benign Lung metastases Introduction Much progress has been made in cancer prevention, early detection, and treatment, with resulting reduction in mortality rates and improved survival for patients with breast cancer [1]. Chest radiography (CXR) is commonly used in newly diagnosed patients with breast cancer for baseline staging [2 6]. For patients without clinical signs of tumor spread, CXR is not recommended, because the frequency of lung metastases is as low as 0.1 % in clinical stage 1 and 0.2 % in stage 2 compared with 1.7 % in stage 3 disease [3]. Recently, chest computed tomography (CT) has been used increasingly for routine staging of breast cancer patients [6 11], and some patients with breast cancer require chest CT scans for other diagnostic purposes such as lung or cardiac disease. Noncalcified solitary or multiple lung nodules are encountered on many of these scans; however, a large majority of these nodules are benign [7, 8], though the likelihood of such nodules representing metastatic disease is uncertain in individual cases. Therefore, we evaluated the size and number of lung nodules on the chest CT scans of breast cancer patients from 1 year prior to breast cancer diagnosis through treatment and follow-up and correlated these findings with breast cancer cell grade and breast cancer clinical stage to provide a basis for determining the probability of malignancy in individual cases.

2 266 Breast Cancer Res Treat (2016) 159: Patient selection Institutional review board approval was obtained, and the requirement for informed patient consent was waived for this retrospective study, which was compliant with the Health Insurance Portability and Accountability Act (HIPAA). The clinical information of all 3313 patients who had been entered into the University of Chicago Medical Center breast cancer registry between January 2002 and December 2011 was reviewed by an experienced chest radiologist (FL, clinical experience more than 15 years). Among these 3313 patients, 100 patients (7 %) had more than one primary breast lesion (96 patients had two breast cancers and four patients had three breast cancers in one or both breasts); for these patients, only the most recently diagnosed primary breast cancer was used for data analysis. Among these patients, 1941 (59 %) patients had no chest CT scan, and 47 (1.4 %) patients had all their CT scans performed more than 1 year prior to breast cancer diagnosis. The remaining 1325 (40 %) patients had 4889 chest CT scans, of which at least one scan had been performed during the period from more than 1 year prior to the breast cancer diagnosis through July 2013; these were the cases used in the current study. Lung nodule inventory The same radiologist reviewed all CT images along with radiology reports and pathology reports, when available, for the 1325 breast cancer patients with chest CT scans. Among these patients, 522 (39 %) patients had only one CT scan, and 803 (61 %) patients had more than one CT scan. All 4889 chest CT scans from these 1325 patients were reviewed to determine whether the patient had noncalcified lung nodules. Among these patients, the criterion for inclusion in this study was the presence of noncalcified lung nodules visible on a CT scan acquired from 1 year prior to the diagnosis of breast cancer through follow-up; the follow-up time period for this study ended in July Only one CT scan per patient was used for nodule analysis (i.e., the scan on which the noncalcified lung nodules were first visible). All chest CT scans and images from other modalities (e.g., CXR and PET) or pathology information, if available, were used to confirm the final diagnosis of the lung nodules. The malignant lung nodules included pathology-proven lung cancers, nodular metastases with pathological proof or clinical proof (increased nodule number and size or response to cancer treatment by follow-up imaging findings), and suspicious or indeterminate nodules according to imaging findings. The benign lung nodule category included both definitively benign nodules (pathology-proven, resolved at follow-up, solid nodules stable over at least 2 years, and nodules with benign patterns of calcifications or fat components) and almost certainly benign nodules based on other benign-appearing imaging features. All noncalcified lung nodules on the CT scans were noted and correlated with the final diagnosis. Data analysis The size of each identified nodule was manually measured as the average of the longest long-axis diameter and the longest short-axis diameter in any one CT section. The size of the largest nodule and the number of all noncalcified nodules with size 2 49 mm were correlated with the final diagnosis. The features of malignancy and benignancy were compared in terms of small (\10 mm) versus large (C10 mm) nodules and solitary versus multiple nodules. The presence of nodular metastases from breast cancer on the chest CT scans was correlated with breast cancer cell grade and clinical stage. A Chi square test for independence was used to compare the benign nodules and the malignant nodules and to compare metastatic nodules from breast cancer and nodular lung cancers based on (1) size (small vs. large) and (2) number (solitary vs. multiple). The Chi square test was also used to compare the patients with lung metastases from breast cancer and patients without lung metastases based on (1) breast cancer cell grade (grade 1 2 vs. grade 3) and (2) clinical stage (stage 0 1 vs. stage 2 3). Results Of the 1325 patients, 812 patients (61 %) had a CT scan acquired during the study period with at least one noncalcified lung nodule (Fig. 1). The remaining 513 (39 %) patients had no lung nodules in any CT scan acquired during the study period. From each of the 812 patients, the first (i.e., earliest) scan in which noncalcified nodules were retrospectively visible was used for analysis. The malignant nodule was used for the analysis if a scan had both malignant and benign nodules. CT section thickness for these 812 CT scans was 1 mm in 580 (71 %) scans, 2 3 mm in 35 (4 %) scans, 4 5 mm in 181 (22 %) scans, and 7 10 mm in 16 (2 %) scans. The CT scan was acquired less than 1 year prior to the diagnosis of breast cancer in 50 (6.2 %) patients, less than 1 year after the diagnosis of breast cancer in 397 (49 %) patients, and more than 1 year after the diagnosis of breast cancer in the remaining 365 (45 %) patients. The recorded clinical information indicated that 637 (79 %) scans had been acquired primarily for breast cancer staging, 70 (9 %) had been acquired for

3 Breast Cancer Res Treat (2016) 159: Fig. 1 Flow diagram of 1325 breast cancer patients with/without lung nodules on chest CT scans and final diagnosis of 812 patients with lung nodules staging or searching for other cancers, 64 (7.9 %) had been requested to evaluate for lung or cardiac disease, and 41 (5.0 %) had been obtained at outside hospitals without available clinical indication. The final diagnoses of the 812 breast cancer patients with at least one noncalcified lung nodule on a CT scan are detailed in Fig. 1. Among the 812 patients, ten patients with a benign nodule and 27 patients with a malignant nodule also had another benign or malignant nodule that was not included in the data analysis. Considering only the largest noncalcified lung nodule, the final diagnosis included 482 (59 %) benign nodules (225 definitely benign and 257 almost certainly benign lesions) and 330 (41 %) malignant nodules (26 primary lung cancers, 19 suspicious malignant, 41 indeterminate lesions, and 244 nodular metastases). The nodule size characteristics of the 812 breast cancer patients with noncalcified lung nodules on a chest CT scan are summarized in Table 1. Among the 812 patients, 615 (76 %) had small (\10 mm) nodules and 197 (24 %) had large nodules (C10 mm). Smaller nodule size was associated with decreased likelihood of malignancy, except in the mm nodule size group. The percentage of scans with malignant nodules was smaller for scans with mm nodules (67 %) than for scans with mm nodules (79 %) or mm nodules (90 %). Large nodules were more often malignant than were small nodules (80 vs. 28 %, P \ 0.001). Multiplicity of nodules was associated with the size of the largest nodule in the 812 breast cancer patients (Table 2). Among the 812 patients, 226 (28 %) had solitary and 586 (72 %) had multiple nodules. A smaller number of nodules were associated with decreased likelihood of malignancy. In very small nodules (2 4 mm), the malignancy rate of solitary and multiple nodules was 8 and 20 %, respectively. Among 54 patients with 2 4 mm malignant nodules, 36 (67 %) patients had nodular metastases (29 from breast cancer) and 18 (33 %) patients had indeterminate nodules. Multiple nodules (46 %) compared with solitary nodules (26 %) at any size were more often malignant than benign (P \ 0.001). Figure 2 shows a small perifissural nodule in the right lung with no change on follow-up CT scans over 2 years, consistent with an intrapulmonary lymph node. Figure 3 shows metastatic nodules from breast cancer in the right lung with increased size and number on follow-up CT scans. A total of 212 (26 %) of the 812 patients with lung nodules (16 % of all 1325 patients) had lung metastases from breast cancer. Another 32 (3.9 %) of these 812 patients had nodular metastases due to other advanced cancers (nine lung cancers, six gynecologic cancers, six urologic cancers, and 11 cancers of other types). Also, 26 (3.2 %) of these patients had pathology-proven primary lung cancer (mostly adenocarcinoma or nonsmall cell carcinoma). Among patients with small nodules, the percentages of patients with primary nodular lung cancer and nodular metastases from breast cancer were 12 and 49 %, respectively (P = 0.02). Among patients with solitary

4 268 Breast Cancer Res Treat (2016) 159: Table 1 Imaging size characteristics of noncalcified lung nodules on chest CT scans in patients with breast cancer Size of largest nodule (mm) malignant nodules benign nodules Percentage of malignant in patient group (%) Chi square P value 2 4 (N = 345) (N = 270) (N = 111) (N = 41) (N = 33) (N = 12) Small size \10 (N = 615) \0.001 Large size C10 (N = 197) Total (N = 812) Table 2 Nodule numbers compared with largest nodule size on chest CT scans in patients with breast cancer Nodule number (mm) malignant nodules benign nodules Percentage of malignant in patient group (%) Chi square P value Solitary nodule (N = 226) (N = ) (N = 65) mm (N = 38) Multiple 2 6 nodules (N = 372) Multiple [6 nodules (N = 214) (N = 222) (N = 205) (N = 159) Solitary nodule (N = 226) \0.001 Multiple nodules (N = 586) Total (N = 812) nodules across all sizes, the percentages of patients with primary nodular lung cancer and nodular metastases from breast cancer were 46 and 12 %, respectively (P = 0.001). No breast cancer cell grade was specified for 205 (15 %) of all 1325 breast cancer patients with chest CT scans, including 33 (16 %) of the 212 patients with lung metastases from breast cancer. Only seven of the 1325 patients had specified cell grade 4; none of these seven patients was among the 212 patients with lung metastases from breast cancer. These cell grade 4 patients were not used for statistical analysis because this patient group was too small. No clinical stage was specified for 380 (29 %) of all 1325 breast cancer patients with chest CT scans, including 66 (31 %) of the 212 patients with lung metastases from breast cancer. Stage 4 breast cancer was recorded for 168 patients, including 53 of the 212 patients with lung metastases from breast cancer. The rate of lung metastases in patients who had stage 4 breast cancer was 32 % (53/168); the CT scan was acquired less than 1 year prior to the diagnosis of breast cancer in two (3.8 %) patients, less than 1 year after the diagnosis of breast cancer in 43 (81 %) patients, and more than 1 year after the diagnosis of breast cancer in the remaining eight (15 %) patients. Analysis by breast cancer cell grade 1 3 and clinical stage 0 3 in patients with lung metastases on chest CT scans is shown in Table 3. Among the 1113 patients with recorded breast cancer cell grade 1 3, 179 (16 %) patients had lung metastases from breast cancer; the CT scan was acquired less than 1 year prior to the diagnosis of breast cancer in three (1.7 %) patients, less than 1 year after the diagnosis of breast cancer in 67 (37 %) patients, and more than 1 year after the diagnosis of breast cancer in the remaining 109 (61 %) patients. Among the 777 patients with recorded clinical stage 0 3, 93 (12 %) patients had lung metastases from breast cancer; the CT scan was acquired less than 1 year prior to the diagnosis of breast cancer in two (2.2 %) patients, less than 1 year after the diagnosis of breast cancer in 20 (22 %) patients, and more than 1 year after the diagnosis of breast cancer in the remaining 71 (76 %) patients. In further analysis, lung

5 Breast Cancer Res Treat (2016) 159: Fig year-old woman with breast cancer (breast cancer cell grade 2 and clinical stage 2B) a Initial chest CT scan shows a perifissural small nodule in the right lung (arrow) and b Follow-up chest CT scan at 38 months shows stable nodule (probable lymph node) after more than 2 years Fig year-old woman with breast cancer (breast cancer cell grade 3 and clinical stage 2B) a Initial chest CT scan shows two small noncalcified lung nodules in the right lung (arrows) and b Follow-up metastases from breast cancer were found more often with breast cancer cell grade 3 (22 %) than with grade 1 2 (10 %) (P \ 0.001) and also more often when patients were clinical stage 2 3 (14 %) than stage 0 1 (7.9 %) (P = 0.03). Discussion CT screening has been used in the early detection of lung cancers, and these scans detect many noncalcified indeterminate lung nodules; however, only a small proportion of these lung nodules are lung cancers [12, 13]. Based on experience from CT screening for lung cancer, guidelines [14] have been proposed by the Fleischner Society for chest CT scan at 26 months shows multiple nodular metastases from breast cancer management of lung nodules detected on routine CT scans in clinical practice, but these guidelines specifically exclude patients with known malignancy, such as breast cancer. One study previously reported the clinical significance of CT-detected silent lung nodules in patients with early breast cancer [7], and another study reported on the value of a preoperative staging CT scan to detect asymptomatic lung and liver metastases in patients with breast cancer [8]. The resulting analyses in these studies indicated that malignancy rates were less than 20 %, even when including relatively large nodules greater than 10 mm [8]. The current study provides a more detailed analysis (imaging findings, final nodule diagnosis, breast cancer cell grade, and clinical stage) for all malignant nodules (including true breast cancer metastases) and benign nodules

6 270 Breast Cancer Res Treat (2016) 159: Table 3 Analysis for breast cancer cell grade and clinical stage in patients with lung metastases from breast cancer on chest CT scans Patient group No. of patients with lung metastases Percentage of malignant in patient group (%) Chi square P value Breast cancer cell grade 1 3 (N = 1113) Grade 1 2 (N = 541) \0.001 Grade 3 (N = 572) Clinical stage 0 3 (N = 777) Stage 0 1 (N = 278) Stage 2 3 (N = 499) detected by chest CT scans in a large series of patients with proven breast cancer. A previous study that addressed the role of surgery in the management of solitary lung nodules in patients with breast cancer reported that among 79 consecutive patients during a 13-year period: 38 (48 %) patients had primary lung cancer, 27 (34 %) patients had metastatic nodules from breast cancer, and 14 (18 %) patients had benign nodules [15]. Surgical series such as this, however, are inevitably biased toward larger and more suspicious nodules that are judged sufficiently suspicious to merit biopsy or resection. Generally, patients with lung metastases have a much higher frequency of multiple masses/nodules than patients with primary lung cancer [16], and the differentiation of a primary lung cancer from solitary metastatic nodules can be very difficult based on imaging findings alone [17]. Based on both biopsy results and imaging findings, the current study indicated that (1) solitary nodules were more often primary lung cancers (46 %) than breast cancer metastases (12 %) and (2) small nodules were more often metastases (49 %) than primary cancers (12 %). Among the 330 (41 %) patients with malignant lung nodules on CT scans in this series, 212 (26 %) patients had nodular lung metastases from breast cancer (the largest group) and 32 (3.9 %) had lung metastases from other cancers, including lung cancer. Also, 26 (3.2 %) patients had pathology-proven primary lung cancer, which is a lower rate than that found in high-risk smokers enrolled in lung cancer screening programs with a range of % among persons with nodules [18 20]. The remaining 60 (7.4 %) patients had indeterminate or suspicious malignant lesions. The malignant nodules among these different categories were not separated for purposes of analysis in this study, although the treatment plans for the primary and metastatic nodules would be different. Kim et al. [8] reported on stage 1 3 breast cancer patients with lung nodules on staging chest CT scans. Of 163 patients with small nodules\10 mm, the nodules of 19 (12 %) patients were nodular metastases from breast cancer, whereas of 17 patients with large nodules (C10 mm), the nodules of three (18 %) patients were metastases from breast cancer, indicating no significant difference in the proportion of breast cancer metastasis among nodules in these two size categories. Lee et al. [7] reported that among 802 breast cancer patients with staging chest CT scans, 34 (4.2 %) patients had indeterminate lung nodules. The follow-up results showed that among 26 patients with \10 mm nodules, six of seven (86 %) patients with solitary nodules and 17 of 19 (89 %) patients with multiple nodules had benign nodules, whereas among eight patients with C10 mm multiple nodules, all nodules were malignant. The current study showed that among patients with multiple nodules greater than 10 mm, 83 % had malignant nodules, most of which were lung metastases from breast cancer; however, the majority of small lung nodules were benign, and in the case of very small nodules (2 4 mm), the malignancy rate of solitary nodules was only 8 %. Although the results indicate that decreased nodule size and increased nodule number are related to an increased likelihood of benignancy, the percentage (67 %) of malignant nodules in the mm size range was lower than among nodules in the mm size range (79 %) or the mm size range (90 %). Imaging findings suggest that nodules in this size range probably were caused by inflammatory changes during and after chemo or radiation therapy. Chest CT scans can upstage 6 % of stage 3 patients to stage 4 [8], and a metastatic workup, including chest CT, is only considered indicated for N2/N3 breast cancer patients with T3 or T4 primary lesions [10]. In the current study, 16 % of patients with breast cancer cell grade 1 3, and 12 % of patients in clinical stage 0 3 had nodular lung metastases from breast cancer. For most of these patients with lung metastases, the metastatic nodules were not identified in CT exams until [1 year after the diagnosis of breast cancer. Lung metastases from breast cancer were found more often with breast cancer cell grade 3 than grade

7 Breast Cancer Res Treat (2016) 159: and when patients were clinical stage 2 3 rather than stage 0 1. Limitations in this retrospective study included that (1) the time period between CT and diagnosis for breast cancer was not consistent, and (2) imaging techniques for chest CT scans were variable. However, even with these limitations, the current study could determine significance of noncalcified lung nodules on chest CT scans of a large number of patients with breast cancer in routine clinical practice; the imaging findings should help oncologists in managing their concern for metastases to reduce unnecessary patient anxiety. In conclusion, lung metastases were especially likely in patients with multiple nodules greater than 10 mm in size, whereas the large majority of solitary nodules and very small multiple nodules on chest CT scans were benign in breast cancer patients. Increased breast cell grade and clinical stage increased the likelihood of lung metastases in these patients. Acknowledgments Funding This work was supported in part, by a Science Council Research Seed Funding Grant from the American Association of Physicists in Medicine. Compliance with ethical standards Disclosure SG. Armato is a consultant for Aduro Biotech, Inc. M.L. Giger is a shareholder of Hologic and Quantitative Insights. H. MacMahon is a shareholder of Hologic and a consultant for Riverain Technologies. F. Li, S.G. Armato, M.L. Giger, and H. MacMahon receive royalties from multiple companies through The University of Chicago (UC Tech). References 1. Jemal A, Siegel R, Ward E, Hao Y, Xu J, Tun MJ (2009) Cancer statistics, CA Cancer J Clin 59: Chen EA, Carlson GA, Coughlin BF, Reed WP, Garb JJL, Frank JL (2000) Routine chest roentgenography is unnecessary in the work-up of stage I and II breast cancer. J Clin Oncol 18: Myers RE, Johnston M, Pritchard K, Levine M, Oliver T, the breast cancer disease site group of the cancer care Ontario practice guidelines initiative (2001) Baseline staging tests in primary breast cancer: a practice guideline. CMAJ 164: Gerber B, Seitz E, Müller H, Krause A, Reimer T, Kundt G, Friese K (2003) 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 82: Puglisi F, Follador A, Minisini AM, Cardellino GG, Russo S, Andreetta C, Terlizzi SD, Piga A (2005) Baseline staging tests after a new diagnosis of breast cancer: further evidence of their limited indications. Annals Oncol 16: Debald M, Wolfgarten M, Kreklau P, Abramian A, Kaiser C, Höller T, Leutner C, Keyver-Paik MD, Braun M, Kuhn W (2014) Staging of primary breast cancer is not indicated in asymptomatic patients with early tumor stages. Oncol Res Treat 37: Lee B, Lim A, Lalvani A, Descamps MJL, Leonard R, Nallamala S, Lewis JS (2008) The clinical significance of radiologically detected silent pulmonary nodules in early breast cancer. Ann Oncol 19: Kim H, Han W, Moon HG, Min J, Ahn SK, Kim TY, Im SA, Oh DY, Han SW, Chie EK, Ha SW, Noh DY (2011) The value of preoperative staging chest computed tomography to detect asymptomatic lung and liver metastases in patients with primary breast carcinoma. Breast Cancer Res Treat 126: Brennan ME, Houssami N (2012) Evaluation of the evidence on staging imaging for detection of asymptomatic distant metastases in newly diagnosed breast cancer. The Breast 21: Chu QD, Henderson A, Kim RH, Miller JK, Burton G, Ampil F, Li BDL (2012) Should a routine metastatic workup be performed for all patients with pathologic N2/N3 breast cancer? J Am Coll Surg 214: James JJ, McMahon MA, Tennant SL, Cornford EJ (2012) CT staging for breast cancer patients with poor prognostic tumours. The Breast 21: Henschke CI, Yankelevitz DF, Mirtcheva R, McGuinness G, McCauley D, Miettinen OS (2002) CT screening for lung cancer: frequency and significance of part-solid and nonsolid nodules. AJR 178: Li F, Sone S, Abe H, MacMahon H, Doi K (2004) Malignant versus benign nodules at CT screening for lung cancer: comparison of thin-section CT findings1. Radiology 233(3): MacMahon H, Austin JHM, Gamsu G, Herold CJ, Jett JR, Naldlch DP, Patz EF Jr, Swensen SJ (2005) Guidelines for management of small pulmonary nodules detected on CT scans: a statement from the Fleischner Society. Radiology 237: Rena O, Rapalia E, Ruffini E, Filosso PL, Oliaro A, Maggi G, Casadio C (2007) The role of surgery in the management of solitary pulmonary nodule in breast cancer patients. EJSO 33: Khokhar S, Vickers A, Moore MS, Mironov S, Stover DE, Feinstein MB (2006) Significance of non-calcified pulmonary nodules in patients with extrapulmonary cancers. Thorax 61: Quint LE, Park CH, Lannettonl MD (2000) Solitary pulmonary nodules in patients with extrapulmonary neoplasms. Radiology 217: New York Early Lung Cancer Action Project Investigators (2007) CT screening for lung cancer: diagnoses resulting from the New York early lung cancer action project. Radiology 243: The National Lung Screening Trial Research Team (2013) Results of initial low-dose computed tomographic screening for lung cancer. N Engl J Med 368: McWilliams A, Tammemagi MC, Mayo JR et al (2013) Probability of cancer in pulmonary nodules detected on first screening CT. N Engl J Med 369: