Selection of colon cancer patients for neoadjuvant

Scandinavian Journal of Gastroenterology. 2014; 49: 202–208

ORIGINAL ARTICLE

Scand J Gastroenterol Downloaded from informahealthcare.com by University of Southern Denmark on 01/20/14 For personal use only.

Selection of colon cancer patients for neoadjuvant chemotherapy by preoperative CT scan ANNE NØRGAARD1, CLAUS DAM1, ANDERS JAKOBSEN2, JOHN PLØEN2, JAN LINDEBJERG3 & SØREN R. RAFAELSEN1 1

Department of Radiology, Vejle Hospital, Kabbeltoft 25, DK-7100 Vejle, Denmark, 2Department of Oncology, Vejle Hospital, Kabbeltoft 25, DK-7100 Vejle, Denmark, and 3Department of Pathology, Vejle Hospital, Kabbeltoft 25, DK-7100 Vejle, Denmark

Abstract Objective: Preoperative staging is essential to plan correct treatment of colon cancer and calls for objective, accurate methods for the introduction of neoadjuvant chemotherapy, which represents a new treatment option. Purpose: To evaluate the diagnostic accuracy of multislice computed tomography (CT) in local staging of colon cancer correlated with histopathological parameters, including criteria for adjuvant chemotherapy. Material and methods. A total of 74 included patients had preoperative CT scans and surgical resection of their colon tumors. Tumor stage (T-stage), extramural tumor invasion (ETI), nodal stage (N-stage), extramural venous invasion (EVI) and the distance from tumor to nearest retroperitoneal fascia (DRF) were retrospectively assessed on the CT scan and compared blindly with the results of the pathological examination, including evaluation of the criteria for adjuvant chemotherapy. Advanced tumors were defined as T3 with ETI ‡5 mm or T4. Results. Sixty-nine percent of the tumors were correctly T-staged by CT, 7% were overstaged and 24% were understaged. As to correct recognition of ETI on the CT scan, the observer was 73% accurate compared with histology (70% sensitivity (95% CI: 53–82%), 78% specificity (95% CI: 60–90%), 81% positive predictive value (PPV) (95% CI: 63–91%) and 66% negative predictive value (NPV) (95% CI: 49–80%). N-stage, EVI and DRF had poor accuracy: 53%, 53% and 64%. All patients with advanced tumors on CT fulfilled the criteria for adjuvant chemotherapy. Positive predictive value: 100% (95% CI: 88–100%). Conclusion. CT has a potential in the preoperative selection of advanced tumors suitable for neoadjuvant chemotherapy without overtreatment of low-risk patients.

Key Words: colonic cancer, CT-spiral, large bowel, sensitivity and specificity, staging

Introduction Colorectal cancer is the second most common type of cancer in developed countries, and is related to significant morbidity and mortality [1]. Colonic cancer will also, in the next few years, represent a considerable burden to the health system with many lost years. Progress in treatment is therefore of utmost importance. Surgery is the cornerstone in the treatment approach and around 80% of the patients are primarily operated [2]. Adjuvant chemotherapy is offered to

patients with high risk of recurrence fulfilling the generally accepted criteria, but the treatment is not too effective, even with combination chemotherapy, and it has only been beneficial to a small subgroup of patients [3,4]. Consequently, the mortality is still high in patients with a locally advanced tumor and new options are called for. Neoadjuvant chemotherapy is now standard in a number of malignant tumors, including breast, rectal, esophageal and ventricular cancer, but it has not been introduced in colon cancer [5–7]. One major reason is the lack of suitable methods for the selection of

Correspondece: Claus Dam, MD, Department of Radiology, Vejle Hospital, Kabbeltoft 25, DK-7100 Vejle, Denmark. Tel: +45 79 40 53 43. E-mail: [email protected]

(Received 28 September 2013; revised 30 October 2013; accepted 30 October 2013) ISSN 0036-5521 print/ISSN 1502-7708 online 2014 Informa Healthcare DOI: 10.3109/00365521.2013.862294


CT selection of colon cancer for neoadjuvant chemotherapy patients. It seems reasonable to assume that patients fulfilling the criteria for adjuvant chemotherapy would be obvious candidates for neoadjuvant treatment, but so far, there are no generally accepted methods of identifying this group of patients before the operation. It is likely that tumors infiltrating other organs or with considerable extramural tumor invasion represent patients with a high risk of recurrence, but the possible correlation between this category and patients fulfilling the criteria for adjuvant chemotherapy has not been presented. Modern CT scans with adequate techniques represent a major step forward and neoadjuvant treatment could possibly be based on CT findings [3,8–10]. The aim of the present study was to investigate the accuracy of preoperative CT scan in identifying highrisk patients in need of adjuvant chemotherapy. Material and methods Eighty-six consecutive patients planned for elective resection of a colon carcinoma in a 22-month period were retrospectively reviewed in a single center setting. The patient data were retrieved in the specific Radiology Information System (RIS)-Picture Archiving and Communications System (PACS) archive at our department. The CT scans of the thorax and abdomen were performed applying a 64-slice CT system (Somatom Sensation, Siemens, Erlangen, Germany) using the following parameters: CARE-Dose: on, 120 kV, up to 200 mA, rotation time 0.5 sec., pitch 1.4, collimation, 0.6 mm, increment 2. Effective dose: approximately 8–9 mSv. Images were obtained with breath-hold technique after intravenous injection of 100 ml iomeprol (Iomeron, Bracco, Milan, Italy) 300 mg I/ml using an automatic injector OptiVantage DH (Mallinckrodt, Cincinnati,

203

Ohio, USA) at an injection rate of 4 ml/sec. Patients with a body weight above 80 kg had 150 ml iomeprol injected at the same injection rate. Care bolus technique with a trigger limit of 100 HU and a delay of 25 s was used. Contrast-enhanced CT scan in the portal phase was performed and interpreted blindly from the histopathologic findings, original clinical CT radiology report and any preceding multidisciplinary colorectal conference discussions. All images were evaluated by a gastrointestinal radiologist with more than 10 years’ experience using an Impax PACS workstation (Agfa, Mortsel, Belgium) with two Coronis monitors (1600 1200 pixels), Megapixels Diagnostic Display System (Barco, Kortijk, Belgium). T-stage assessed by CT was evaluated according to the TNM system. ETI was defined as maximal tumor outgrowth from the intestinal wall in millimeters. The shortest distance from tumor to nearest retroperitoneal fascia (DRF) was noted. Also, N-stage was noted as well as number of lymph nodes and the size of the largest lymph node in millimeters. Lymph nodes larger than 5 mm were considered metastatic. Extramural venous invasion (EVI) detected on CT was noted. Patients were stratified into good prognosis and poor prognosis groups based on the findings on their staging CT examination. The criteria for poor prognosis were categorized as T4 tumor or T3 tumor with CTETI ‡5 mm as evaluated on the axial CT images (Figure 1). All patients had a resection within two weeks after the CT scan. The resected specimens were transferred to the pathology department and fixed for two days in neutral buffered formalin. After fixation, all tumors were sliced transversally. The deepest penetration from the outer edge of tunica muscularis, the minimal distance from the invasive front to the

202

203

W 476 : L 50

W 400 : L 50

Figure 1. (A) Large T4, N1, M0 cecal tumor with tumor outgrowth into the small intestine. (B) T3, N0, M0 colonic tumor with a 10 mm tumor outgrowth close to the anterior abdominal wall (arrow).

204

A. Nørgaard et al.

Patients with primary CRC examined by MDCT 86


Not detected by MDCT pT1: 1 pT3: 6 pT4: 2

Detected by MDCT but without reference standard

Detected by MDCT and with reference standard

Palliative treatment: 2 Duplex tumors: 1

Study population: 74

Figure 2. Patient flow chart of the 86 patients referred to preoperative examinations. The study population included 74 patients with colon cancer.

non-peritonealised margin and the minimal distance from the peritoneal surface were measured on the slices. These distances were confirmed microscopically. The resected specimens were classified according to the pTNM system by an experienced gastrointestinal pathologist. The peritoneal surface was considered involved (pT4) if viable cancer was detected outside the peritoneal lining or infiltrating adjacent organs. The pathologist was blinded to the CT findings. The study was approved by the National Data Protection Agency, 2 April 2009. Statistical analyses were performed using NCSS software, Kaysville, UT, USA. The statistics were descriptive. Sensitivity, specificity, positive and negative predictive value with 95% confidence intervals were calculated. A p Value less than 0.05 was considered to indicate a statistically significant difference.

T-stage Pathologic examination found two pT1, three pT2, 51 pT3 and 18 pT4 tumor stages. CT examination correctly identified no pT1 (0%), three pT2 (100%), 41 pT3 (80%) and seven pT4 (39%) lesions. CT diagnosis overestimated two lesions in the pT1 group, no lesions in the pT2 group and three lesions in the pT3 group. In total, five patients (7%) were overstaged. CT diagnosis underestimated no lesions in the pT2 group, seven lesions in the pT3 group and 11 lesions in the pT4 group. Totally, 18 patients (24%) were understaged (Table I). Twelve patients were staged M1 with distant metastases. Twenty-eight patients had a poorly differentiated tumor and 45 patients had a well or moderately well differentiated tumor. One patient’s tumor differentiation was not assessed (missing data). There were five T1/T2 tumors and they were all well/moderately well differentiated tumors. Extramural tumor invasion

Results Out of 86 patients, 12 were excluded: nine patients had tumors that could not be identified on the CT scan (pT1 = 1, pT3 = 6, pT4 = 2; out of these five were located in the sigmoid colon, two in the cecum, one in the ascending colon and one in the transverse colon). Two patients were inoperable and one patient had two tumors in the sigmoid colon located close to each other. On the CT scan, it was only possible to see one of the two tumors, and further, it was not possible to characterize it (Figure 2). The remaining 74 patients with primary colonic cancer represented the study population; 29 males and 45 females with a mean age of 72 years, range: 32–90 years.

All patients with pT4 tumor and 47% of pT3 tumors were advanced with a pETI of 5 mm or more. In the correct recognition of CTETI on MDCT, the observer was 73% accurate compared with histology (70% sensitivity (95% CI: 53–82%), 78% specificity Table I. T-stage. CT examination correlated with histopathologic pT-stage.

CTT1 CTT2 CTT3 CTT4

Total

pT1

pT2

pT3

pT4

Total

0 1 1 0 2

0 3 0 0 3

0 7 41 3 51

0 1 10 7 18

0 12 52 10 74

205

CT selection of colon cancer for neoadjuvant chemotherapy Table II. Cross tabulation of outgrowth measurement results by CT and histopathology in all 74 patients. Advanced tumors were defined as T4 tumor or T3 tumor with ETI ‡5 mm. Outgrowth at histopathology Radiology (CT) Early (