Correlation between cancer antigen 15.3 value and qualitative and semiquantitative parameters of positron emission tomography/computed tomography in breast cancer patients.

Abstract

The association of PET/CT and tumor markers can be considered complementary, since any significant increases of tumor markers can indicate the presence of disease while PET/CT is able to detect and describe the tumor sites. In this retrospective, single-institution study, we determine the correlation between cancer antigen (CA) 15.3 value and qualitative and semi-quantitative PET/CT data in breast cancer (BC) patients. 193 BC patients (median age 61 yrs) already treated with primary treatment (surgery and others) were identified through institutional databases. All patients underwent PET/CT for increase in tumor markers, post-therapy evaluation, restaging and doubtful conventional imaging for disease relapse. The CA15.3 values before PET/CT scan were collected for all patients. Clinical outcome was defined as presence or absence of disease recurrence based on follow-up data (histological or imaging findings). CA15.3 quartile values and qualitative and semi-quantitative (maximum Standardized Uptake Value - SUVmax) PET/CT findings were compared with chi-square test and linear regression analysis. The mean value of CA15.3 was significantly higher in patients with positive than negative PET/CT (67.51±120.92 vs. 25.54±17.54, p<0.005). PET/CT was positive in 107 (55%) and negative in 86 (45%) patients; CA15.3 value was considered abnormal (≥ 31 UI/mL) in 85 (44%) patients; 57 of them showed positive PET/CT while 28 a negative scan (67 vs. 33%, p<0.05). In all 193 patients, the disease recurrence was found in 71 (37%), whereas 122 (63%) were disease-free. The diagnostic accuracy of PET/CT in all 193 patients was 74%. Among patients with normal CA15.3 value (n=108), 50 showed positive PET/CT; 24 out of these latter 50 patients (48%) had recurrence of disease. The combination of the highest quartile of CA15.3 (value>45 UI/ml) and FDG PET/CT determined high sensitivity and accuracy (92% and 82%, respectively) but a low specificity (50%) for restaging BC patients. The highest specificity (~ 70%) was found when PET/CT and 2nd quartile of CA15.3 (value: 12.95-25) were associated. No correlation between CA15.3 values and SUVmax was found (p=0.489); whereas a trend in increase of the CA15.3 value and SUVmax in the presence of visceral and no-visceral site of disease (22.4±16.2, 64.9±108 and 6.4±4.2, 8.2±5.1, respectively) was identified. The value of CA15.3 and PET/CT findings are consistently complementary. About 25% of BC patients with a negative CA15.3 value had a positive PET/CT and disease relapse. SUVmax and CA15.3 values are not correlated.