World J Surg (2007) 31:1502–1506 DOI 10.1007/s00268-007-9060-0

Clinical Usefulness of Chest Radiography in Detection of Pulmonary Metastases After Curative Resection for Colorectal Cancer Won-Suk Lee Æ Seong Hyeon Yun Æ Ho-Kyung Chun Æ Woo Yong Lee Æ Haeran Yun

Published online: 5 May 2007 Ó Socie´te´ Internationale de Chirurgie 2007

Abstract Purpose The purpose of this study was to evaluate the effectiveness of chest radiography (CXR) and abdominal computed tomography (CT) for detecting pulmonary metastases after curative surgery for colorectal cancer. Methods We performed a retrospective analysis of the records of all patients with pulmonary metastasis from colorectal cancer who underwent curative resection between 1994 and 2004 at our institution. Results Pulmonary metastases were detected in 193 patients by either CXR or abdominal CT. They were initially detected by CXR in 87 patients (45.1%) and by abdominal CT in 106 patients (54.9%). In the CXR group, the patterns of pulmonary recurrence were as follows: solitary (n = 38, 43.7%), multiple unilateral (n = 11, 12.6%), and multiple bilateral (n = 38, 43.7%). In the CT group, there were 22 patients (20.8%) with a solitary nodule, 17 patients (16.0%) with multiple unilateral nodules, and 67 (63.2%) with multiple bilateral nodules. The overall survivals of the CXR group and abdominal CT group were 34.6% and 31.7%, respectively (p = 0.312). There was no difference in the median disease-free interval between the CXR group and the abdominal CT group (23.8 vs. 23.2 months, p = 0.428). Conclusions Although this study is limited by its small sample size, it can be speculated that abdominal CT with lower thorax images may replace CXR in surveillance programs.

W.-S. Lee
S. H. Yun (&)
H.-K. Chun
W. Y. Lee
H. Yun Department of Surgery, Samsung Medical Center, Sungkyunkwan University School of Medicine, 50 Ilwon-dong, Gangnam-gu, Seoul 135-710, Korea e-mail: [email protected]

123

Colorectal cancer (CRC) accounts for 10% to 15% of all cancers; it is the second leading cause of cancer deaths in Western countries [1] and the fourth most common cancer in Korea [2]. Approximately one-fourth of CRC patients have metastatic disease on diagnosis. Three-fourths of patients are able to undergo radical surgery as their main curative treatment. However, more than 40% of these patients have a recurrence during follow-up, mostly during the first 3 years following surgery [3]. The surveillance program aims to enhance the frequency of early diagnosis of recurrent disease that could be treated with curative intent and thus improve survival. Isolated pulmonary metastases account for 0.9% to 10% of first-site recurrences in CRC patients with a high incidence in rectal cancer patients [4–6]. Of these patients, only 10% are candidates for pulmonary resection [4, 6]. Although there is no consensus regarding the prognostic factors in various studies, it is generally accepted that isolated pulmonary metastases benefit from resection, with as high as 60% overall survival [7]. Chest radiography (CXR) has been reported to identify only 1.8% to 12.0% of patients with resectable pulmonary metastasis [4]. Although routine CXR follow-up may enhance the detection rate of pulmonary metastases, it is not uncommon to encounter normal chest radiographs when metastatic pulmonary nodules are seen to be present on abdominal CT (computed tomography) scans. A chest CT scan has been shown to be more sensitive than CXR for identifying a resectable pulmonary recurrence [8]. Reports that focus on methods to detect pulmonary metastases are scarce. Although the American Society of Clinical Oncology (ASCO) guidelines for CRC surveillance after surgery in 2005 discouraged the use of annual chest radiography as surveillance for pulmonary

World J Surg (2007) 31:1502–1506

metastases [9], there are several surveillance guidelines with different recommendations [4–6, 10]. In addition, adherence of surveillance practices to the guidelines may not be perfect, as suggested by a previous study [10]. Thus, we investigated the role of chest radiography in screening for CRC metastases to the lung in a Korean population. We retrospectively reviewed 193 CRC patients who underwent curative resection of their CRC and developed metastatic pulmonary lesions during surveillance. We attempted to delineate the potential benefit of CXR as a surveillance tool.

1503 Table 1 Patient characteristics Characteristic

Data (n = 193)

Sex (M:F)

1.3:1.0

Age (years), mean and range

60.6 (27.5–88.2)

Location Colon

50 (25.9%)

Rectum

143 (74.1%)

Stage I

5 (2.6%)

II

44 (22.8%)

III

144 (74.6%)

Pattern of lung recurrence

Patients and methods Between 1994 and 2004, a total of 3503 patients were treated with potentially curative surgery for CRC at our institution. Surveillance for recurrence following surgery was outlined as follows: A physical examination, serum carcinoembryonic antigen (CEA) assay, CXR, and spiral abdominal CT scan were performed every 6 months for 3 years and annually thereafter. The upper limit of each CT was set 3 cm above the highest point of the diaphragm. Other examinations, such as colonoscopy and abdominal ultrasonography, were selected and performed every 6 to 12 months depending on the status the patient. For the 4 years from February 2000 to December 2004, a positron emission tomography (PET) scan was performed in patients with suspected pulmonary recurrence when the results of CT scans or other imaging modalities, such as CXR, were not conclusive. During the surveillance, pulmonary metastases were detected in 193 patients. Exclusion criteria were as follows: (1) lesion initially diagnosed as stage IV according to the AJCC Cancer Staging Manual, 6th edition [11]; (2) other metastases detected prior to pulmonary metastases; (3) synchronous metastases detected during the first 6 months after curative resection; (4) double primary cancers detected during the follow-up; (5) patients initially evaluated with chest CT or PET scans. A total of 193 patients were included in this study. The following data were collected from medical records of each patient: age, sex, location and initial stage at the time of surgery, location of pulmonary metastases, initial imaging tool for detecting pulmonary metastases. If pulmonary metastases were concurrently documented at CXR and abdominal CT scanning, the imaging tool that detected pulmonary metastases first was recorded as the method of surveillance in this study. Pulmonary metastases were pathologically or cytologically confirmed by fine-needle aspiration (FNA) on surgical specimens. A radiological diagnosis was assigned, including a description of the size and number of lesions.

Single

60 (31.1%)

Unilateral multiple

28 (14.5%)

Bilateral multiple

105 (54.4%)

CEA > 5 ng/ml < 5 ng/ml

91/189 (48.1%) 98/189 (51.9%)

CEA, carcinoembryonic antigen

The primary objective of the study was to delineate the time to the first detection of pulmonary metastases for each imaging modality. Time to the first detection was defined as the time between the date of surgery for CRC and the time of the first detection of pulmonary metastases. The disease-free interval (DFI) after primary cancer resection was estimated by the Kaplan-Meier method. The chi-squared test or Fisher’s exact test was used as appropriate. A level of p < 0.05 was regarded as statistically significant.

Results Patient characteristics Patient characteristics are listed in Table 1. According to the TNM classification proposed by the AJCC Cancer Staging Manual, 6th edition,11 5 CRCs (2.6%) were stage I, 44 CRCs (22.8%) were stage II, and 144 CRCs (74.6%) were stage III. Altogether, 105 patients (54.4%) with pulmonary relapses had bilateral multiple lesions; 160 patients (82.9%) had lower lobe involvement; 33 patients (17.1%) had pulmonary relapse other than in the lower lobe. The CEA level was available for 189 patients. The mean CEA level was 84.8 ng/ml (range 0.1–2250 ng/ml). Altogether, 48.1% of patients had a CEA level > 5 ng/ml. Pulmonary metastases were confirmed by FNA cytology in 22 patients (11.3%); video-assisted thoracoscopic biopsy was performed in 10 patients (5.1%); and surgical resection was done in 56 patients (29.0%). A total of 85% of the

123

1504

World J Surg (2007) 31:1502–1506

Table 2 Pattern of lung metastases according to lung lobes Involved area

CXR (n = 87)

Abdominal CT (n = 106)

Lower lobe

P

Colon (n = 50)

Rectum (n = 143)

0.950

P

Total patients (n = 193)

0.158

Single RLL

11 (12.7%)

12 (11.3%)

6 (12.0%)

17 (11.9%)

23 (11.9%)

LLL

8 (9.2%)

11 (10.4%)

6 (12.0%)

13 (9.1%)

19 (9.8%)

RLL + LLL

7 (8.1%)

20 (18.9%)

8 (16.0%)

19 (13.2%)

27 (13.9%)

RLL + LUL

1 (1.1%)

1 (0.9%)

0

2 (1.4%)

2 (1.0%)

RUL + LLL

0

0

0

0

0

RML+LLL

0

2 (1.9%)

1 (2.0%)

1 (0.7%)

2 (1.0%)

RLL ± LLL + other lobe

34 (39.0%)

53 (50.0%)

22 (44.0%)

65 (45.5%)

87 (45.1%)

26 (29.9%)

7 (6.6%)

7 (14.0%)

26 (18.2%)

33 (17.1%)

Multiple

No lower lobe involvement

0.129

0.229

CXR, chest radiograph; RLL, right lower lobe; LLL, left lower lobe; LUL, left upper lobe; RUL, right upper lobe; RML, right middle lobe

selected patients had completed CXR, and 81.0% of the selected patients had completed abdominal CT at the 6month interval for the first 3 years and annually thereafter. In all, 193 patients were included in this study. Detection method and pattern of pulmonary recurrence In all, 160 patients (67.0%) had lower lobe pulmonary metastases (Table 2). Table 3 provides the pattern of the first initial detection method of pulmonary relapses. In the CXR group, patterns of pulmonary recurrences were as follows: solitary (n = 38, 43.7%), multiple unilateral (n = 11, 12.6%), multiple bilateral (n = 38, 43.7%). In the

CT group, there were 22 patients (20.8%) with a solitary nodule, 17 patients (16.0%) with multiple unilateral nodules, and 67 (63.2%) with multiple bilateral nodules. In both the CT and CXR groups, multiple bilateral pulmonary nodules were detected with the highest frequency compared with single or multiple unilateral pulmonary nodules, with statistical significance (p = 0.003). The 5-year survival rate of the CXR group with pulmonary resection was 56.3%, whereas that for the CT group who underwent surgical resection was 52.8% (p = 0.884). There was no statistically significant difference in median DFI between the CXR group and the abdominal CT group (19.3 vs. 17.3 months, p = 0.608). Diagnostic method according to stage

Table 3 First diagnostic method for detection of pulmonary metastases according to colorectal cancer stage and pattern of recurrence Parameter

CXR (n = 87)

Abdominal CT p (n = 106)

26 (29.9%)

23(21.6%)

Stage I + II III

61 (70.1%)

83 (78.3%)

62.6 (29.0– 88.5)

58.9 (27.1– 81.6)

Solitary

38 (43.7%)

22 (20.8%)

Multiple, unilateral

11 (12.6%)

17 (16.0%)

Multiple, bilateral

38 (43.7%)

67 (63.2%)

33 (37.9%) 54 (62.1%)

23 (21.7%) 83 (78.3%)

Age (years), mean and range

0.407

Pattern of relapse 0.003

Surgery Yes No

0.940

5-Year overall survival Surgery

56.3% (n = 33) 52.8% (n = 23) 0.884

No surgery

19.7% (n = 54) 19.4% (n = 83)

123

The method of detecting the pulmonary metastases was not significantly different according to initial stage of the CRC (p = 0.442). In all, 81 (56.2%) patients with initial stage III had multiple metastases (p = 0.05). CXR revealed abnormal lung lesions in 87 (45.1%) patients. Nevertheless, 66 (34.2%) patients had normal CXR findings despite of the presence of pulmonary metastases on CT. Of 87 patients, 25 (28.3%) had solitary pulmonary metastasis on CXR and chest CT scan, 27 (31.0%) had solitary pulmonary metastasis on CXR and multiple lesions on chest CT scan, and the remaining 35 (40.2%) had multiple pulmonary metastases on both CXR and chest CT scan. Overall survival The overall survival rates of the CXR group and abdominal CT group were 34.6% and 31.7%, respectively (p = 0.312). The median overall survival durations for the CXR and CT groups were 48.2 and 44.3 months, respectively. There was

World J Surg (2007) 31:1502–1506

no difference in median DFI between the CXR and abdominal CT groups (23.8 vs. 23.2 months, p = 0.428) Of the 193 patients, 56 (29.0%) underwent curative pulmonary resection. The 3- and 5-year DFI rates were 41.8% and 34.8%, respectively. The DFI for patients who underwent curative resection ranged from 0.70 to 70.0 months (median 23.4 months). The 3- and 5-year overall survival rates in the pulmonary resection group were 77.1% and 54.9%, respectively. The overall survival duration ranged from 7.6 to 123.9 months (median 68.2 months).

Discussion Most recurrence of metastases from CRC occurs during the first 2 years of follow-up. Liver remains the most common metastatic site, reaching 40% to 80% of all metastases, followed by pulmonary metastasis in 10%, usually in association with other extrathoracic systemic disease [12, 13]. It remains controversial whether patients benefit from early detection of recurrent CRC [14], and screening for these patients poses a challenging dilemma for the surgeon. The DFI may be influenced by the diagnostic methods and the frequency of examinations. In turn, the accuracy of diagnostic methods may be influenced by the time elapsed from the primary operation to the detection of recurrence [15]. Conventional chemotherapy of metastatic CRC with fluorouracil and leucovorin with or without irinotecan or oxaliplatin may prolong progression-free survival and overall survival [16]; unfortunately, the long-term results have been less than satisfactory. Because pulmonary resection for metastatic pulmonary tumor is currently the treatment of choice for patients with colorectal metastasis, it is critical that patients undergo proper surveillance for early detection. There are seven randomized trials and descriptive series from various institutions on surveillance programs for CRCs operated on with curative intent, with the assumption that early detection results in improved survival [6, 8, 17– 21]. Data from these trials are heterogeneous and inconsistent. Figueredo et al. [22] concluded that follow-up programs do improve survival, and these programs should include frequent visits and performance of CEA assays, CXR, liver imaging, and colonoscopy. Desch et al. [9] reported an update of colorectal surveillance in which they stated that yearly CXR is not recommended. Previous studies demonstrated that the 5-year survival rates of patients who underwent resection of pulmonary metastases of CRC varied from 32.4% to 60.0% [7]. The overall 3- and 5-year survival rates for 56 patients who underwent pulmonary resection were 77.1% and 54.9%, respectively. Abdominal CT has identified 23 patients (41.0%, 23/56) in this group of patients, which is compa-

1505

rable to the 45.5% reported by Chau et al. [8]. They also reported that annual follow-up chest CT scans had been shown to be more sensitive than CXR for identifying resectable pulmonary recurrences. Although there was no difference in DFI between the CXR group and the abdominal CT group (23.8 vs. 23.2 months, p = 0.428), our data suggest that waiting 3 years for annual CT of the abdomen and chest after primary therapy in patients who are at high risk of recurrence might be too long, as recommended by ASCO [9]. Needless to say, the main purpose of abdominal CT is to screen for locoregional metastases and liver metastases. It is important to note that the right lower lobe or left lower lobe was involved in 82.9% (n = 160) of all pulmonary metastases, and abdominal CT was able initially to detect 66 patients (34.2%) with normal CXRs. It can be suggested that elevating the upper level of the scan for routine abdominal CT scanning or performing chest CT might result in better detection of pulmonary metastases.

Conclusions Although this study has its limitations due to its retrospective nature, our results suggest that the time to first detection of pulmonary metastases, the rate of complete resection of pulmonary metastases, and DFIs were not significantly different between the CXR and CT groups. Although this study is limited by a small sample size, it can be speculated that abdominal CT with lower thorax images may replace CXR in surveillance programs.

References 1. Jemal A, Murray T, Samuels A, et al. (2003) Cancer statistics, 2003. CA Cancer J Clin 53:5–26 2. Shin HR, Ahn YO (2002) Cancer incidence in Korea. Cancer Res Treat 34:405–408 3. Abulafi AM, Williams NS (1994) Local recurrence of colorectal cancer: the problem, mechanisms, management and adjuvant therapy. Br J Surg 81:7–19 4. Anthony T, Simmang C, Hyman N, et al. (2004) Practice parameters for the surveillance and follow-up of patients with colon and rectal cancer. Dis Colon Rectum 47:807–817 5. Meyerhardt JA, Mayer RJ (2003) Follow-up strategies after curative resection of colorectal cancer. Semin Oncol 30:349–360 6. Ohlsson B, Palsson B (2003) Follow-up after colorectal cancer surgery. Acta Oncol 42:816–826 7. Sakamoto T, Tsubota N, Iwanaga K, et al. (2001) Pulmonary resection for metastases from colorectal cancer. Chest 119:1069– 1072 8. Chau I, Allen MJ, Cunningham D, et al. (2004) The value of routine serum carcino-embryonic antigen measurement and computed tomography in the surveillance of patients after adjuvant chemotherapy for colorectal cancer. J Clin Oncol 22:1420– 1429

123

1506 9. Desch CE, Benson AB 3rd, Somerfield MR, et al. (2005) Colorectal cancer surveillance: 2005 update of an American Society of Clinical Oncology practice guideline. J Clin Oncol 23:8512–8519 10. Boulin M, Lejeune C, Le Teuff G, et al. (2005) Patterns of surveillance practices after curative surgery for colorectal cancer in a French population. Dis Colon Rectum 48:1890–1899 11. Greene FL, Page DL, Fleming ID, et al. (2002) AJCC Cancer Staging Manual. 6th edition. New York, Springer-Verlag 12. Inoue M, Kotake Y, Nakagawa K, et al. (2000) Surgery for pulmonary metastases from colorectal carcinoma. Ann Thorac Surg 70:380–383 13. Moore KH, McCaughan BC (2001) Surgical resection for pulmonary metastases from colorectal cancer. ANZ J Surg 71:143– 146 14. Kronborg O (1994) Optimal follow-up in colorectal cancer patients: what tests and how often? Semin Surg Oncol 10:217– 224 15. Kjeldsen BJ, Kronborg O, Fenger C, et al. (1997) The pattern of recurrent colorectal cancer in a prospective randomised study and the characteristics of diagnostic tests. Int J Colorectal Dis 12:329–334

123

World J Surg (2007) 31:1502–1506 16. Saltz LB, Cox JV, Blanke C, et al. (2000) Irinotecan plus fluorouracil and leucovorin for metastatic colorectal cancer: Irinotecan Study Group. N Engl J Med 343:905–914 17. Kjeldsen BJ, Kronborg O, Fenger C, et al. (1997) A prospective randomized study of follow-up after radical surgery for colorectal cancer. Br J Surg 84:666–669 18. Makela JT, Laitinen SO, Kairaluoma MI (1995) Five-year follow-up after radical surgery for colorectal cancer: results of a prospective randomized trial. Arch Surg 130:1062–1067 19. Pietra N, Sarli L, Costi R, et al. (1998) Role of follow-up in management of local recurrences of colorectal cancer: a prospective, randomized study. Dis Colon Rectum 41:1127–1133 20. Schoemaker D, Black R, Giles L, et al. (1998) Yearly colonoscopy, liver CT, and chest radiography do not influence 5-year survival of colorectal cancer patients. Gastroenterology 114:7–14 21. Secco GB, Fardelli R, Gianquinto D, et al. (2002) Efficacy and cost of risk-adapted follow-up in patients after colorectal cancer surgery: a prospective, randomized and controlled trial. Eur J Surg Oncol 28:418–423 22. Figueredo A, Rumble RB, Maroun J, et al. (2003) Follow-up of patients with curatively resected colorectal cancer: a practice guideline. BMC Cancer 3:26