CA15-3 is a useful serum tumor marker for diagnostic integration of hybrid positron emission tomography with integrated computed tomography during follow-up of breast cancer patients

Breast cancer (BC) is by far the most frequent type of cancer in women worldwide, but it is associated with relatively lower mortality rates, as it ranks fifth in cancer-related deaths overall [1]. These data mostly reflect improvements in the treatment of BC recurrences and metastases due to the availability of new drugs and biological therapies [1]. The efficacy of these medications is greatly improved if recurrent or metastatic disease is detected early [2]. Thus, during recent years, the introduction of 2-deoxy-2-[¹⁸ F] fluoro-d-glucose (¹⁸FDG) positron emission tomography (PET) with integrated computed tomography (CT), known as PETCT, has become a helpful tool for this purpose [3]. Indeed, PETCT is able to detect recurrences at very early stages of development due to the abnormal uptake of ¹⁸FDG by neoplastic cells before the onset of morphologic changes that are detectable with conventional imaging [4, 5]. Whole-body PETCT is a very useful diagnostic technology, especially when BC recurrence is suspected in asymptomatic patients with elevated tumor marker levels and negative conventional imaging results [6‐8]. In this case, the early detection of disease relapse greatly improves the chances for successful treatment. However, despite this important feature, an intense debate remains among scientific organizations about the value of tumor markers in predicting clinical or radiological findings of disease relapse [9]. To date, only the European Group on Tumor Markers (EGTM) suggests that an increase in serum CA15-3 often predicts clinical or radiological signs of disease recurrence in BC patients, with an estimated lead time of 2–9 months [10‐12].

In this study, we performed a retrospective analysis of the medical records of 45 BC patients with suspicion of recurrence who underwent both PETCT scans and serial measurements of CA15-3 to asses if: i) a serial increase in CA15-3 was optimal for recommending addressing PETCT examination of BC patients during follow-up; ii) CA15-3 serum levels could be correlated with PETCT findings in terms of number of detectable FDG positive lesions; iii) elevated CA15-3 serum levels are also predictive of disease relapse in patients who at the end of their therapeutic treatment have a negative PETCT result and are positive for CA15-3.

The daily clinical experiences in monitoring BC patients after primary therapy demonstrate that any significant increase in tumor markers, even in the absence of other clinical or instrumental signs of cancer, justifies PETCT examination [14, 15]. Nevertheless, the usefulness of biomarkers in monitoring BC patients after primary therapy is still the subject of intense debate in the scientific literature [16]. In particular, the current guidelines of the American Society of Clinical Oncology do not recommend measurement of serum biomarkers such as CarcinoEmbryonic Antigen (CEA) and CA15-3 for monitoring BC patients during follow-up [17]. Conversely, the EGTM, in agreement with the National Academy of Clinical Biochemistry guidelines and the European Association for Nuclear Medicine, suggests that increasing levels of serum tumor markers may often precede clinical or radiological signs of disease recurrence [18, 19]. In addition to these recommendations, a growing number of scientific papers have re-evaluated the value of increased CA15-3 serum levels for early detection of recurrence, showing that serum determination of CA15-3 improves the diagnostic accuracy of PETCT [20‐23].

Given these reports, we decided to perform a retrospective study of a group of 45 BC patients who were followed up in our institution with PETCT scans and measurement of CA15-3 serum levels according to their clinical care protocol. Disease relapse was documented by prolonged clinical or radiological follow-up in 16 out of 45 patients (35%). In relapsing patients, the CA15-3 levels were ≥25 UI/mL in 12 out of 16 cases (75% sensitivity), whereas in disease-free patients, CA15-3 levels were <25 UI/mL in 22 out of 29 cases (76% specificity). The CA15-3 positive and negative predictive values were 63% and 84%, respectively. PETCT was able to detect recurrence or metastatic disease in 12 out of 16 patients (75% specificity), whereas in 23 out of 29 (79%) non-relapsing BC cases, PETCT did not show pathological SUV sites. The positive and negative predictive values of PETCT were 67% and 85%, respectively.

For diagnostic integration of the two methods, we correlated the CA15-3 serum levels with PETCT findings and found a significant increase in the median CA15-3 serum levels (p-value <0.05) in patients with multiple metastatic sites (8 out of 12 patients with positive PETCT) compared to those with a single metastatic site (4 out of 12 patients with positive PETCT). Specifically, in relapsing BC patients with multiple metastatic sites, the median value of CA15-3 was 49.1 UI/mL (range 8.2–298.4 UI/mL), whereas in those with a single metastatic site, the median value was 17.4 UI/mL (range 5.8–44.2 UI/mL).These data showed that, according to previous studies, elevated levels of CA15-3 are predictive of disease dissemination rather than local recurrence [24‐26]. In addition, the highest serum levels of CA15-3 (≥120 UI/mL) were found in patients with hepatic involvement in agreement with Nicolini and colleagues [27].

Based on these findings, we asked if CA15-3 provides earlier detection of a relapse in comparison to PETCT. To test this idea according to other studies [22, 26], for each patient, we collected three serial serum CA15-3 measurements from our internal archives. As shown in Figure 1, the serial measurements were performed at 6–9 months (270 ± 102 days), 3–6 months (139 ± 79 days), and 0–3 months (20 ± 30 days) before the PETCT scan. In patients with a positive PETCT scan, the mean CA15-3 value 6–9 months before the PETCT scan was 40.01 ± 42.40 UI/mL (median 23.3 UI/mL, range 6.2–159 UI/mL), 3–6 months before the PETCT scan was 46.17 ± 42.97 UI/mL (median 29.20 UI/mL, range 8–155 UI/mL), and 0–3 months before the PETCT scan was 52.64 ± 68.33 UI/mL (median 31.9 UI/mL, range 5.8–298.4 UI/mL). As expected, in patients with a negative PETCT scan, the mean CA15-3 values 6–9 months, 3–6 months, and 0–3 months before the PETCT scan were <25 UI/mL. In PETCT positive cases the increase in the percentage of CA15-3 between the time spans of 3–6 and 6–9 months was 15%, whereas that between 0–3 and 3–6 months was 14%. Moreover, using ROC curve analysis, we found that the AUC value for CA15-3 was 0.81 for both 0–3 and 3–6 months before the PETCT scan and 0.77 6–9 months before the scan. According to Evangelista et al. [22], these results indicate that an increase in CA15-3 3–6 months (139 ± 79 days) before PETCT could already identify BC patients at risk for relapse, and thus, the biochemical signal of the presence of a tumor could predict positive metabolic imaging of disease. This conclusion is interesting especially in BC patients who undergo PETCT examination while receiving anti-hormonal or chemotherapy medications. In fact, these medications reduce the tumor avidity for FDG and consequently, the sensitivity of PETCT [22]. In our series, 10 BC patients were receiving therapeutic treatment, and PETCT and CA15-3 serum levels were obtained every 6 months. In four of those 10 patients, PETCT was negative and CA15-3 serum levels were higher than 25 UI/mL (45.3, 84.6, 27.8, and 26.2 UI/mL). Nevertheless, they relapsed after a median time of 158 days and showed an increase in the CA15-3 serum levels (116.8, 96.4, 44.2, and 30 UI/mL) and a positive PETCT scan. Conversely, the control group, which was composed of the remaining patients (n = 6) who had a negative PETCT scan and low CA15-3 levels, was free from disease even after 1 year of follow-up. Figure 2 shows an example patient at the end of chemotherapy treatment with a negative PETCT result and a positive CA15-3 value (45.3 UI/mL). This patient relapsed after 6 months and showed positive PETCT findings and CA15-3 levels that had increased 2.5-fold (116 UI/mL). These results show for the first time the usefulness of CA15-3 for identifying patients with false negative PETCT scans that require a stringent follow-up due to a risk for developing disease relapse.

Our study, although performed with a small case series, shows that a serial increase in CA15-3 can predict positive PETCT results and detect the presence of residual disease in patients during therapeutic treatment who show negative PETCT results. The diagnostic integration of CA15-3 with PETCT could be a useful strategy for optimizing clinical protocols for early detection of disease relapse. Indeed, the monitoring of CA15-3 may be useful for correct timing of performing PETCT for early detection of cancer lesions and initiation of appropriate therapeutic intervention. Further studies are necessary to identify additional serum biomarkers that can be integrated with modern molecular imaging instruments. Indeed, CA15-3 is not useful in cases of local regional recurrences or for screening purposes. Innovative biomarkers such as microRNAs and circulating tumor DNA are promising, however, they need to be validated and translated into clinical practice.