Clinical impact of follow‐up imaging on mortality in Korean breast cancer patients: A national cohort study

Abstract Background As the incidence of breast cancer has increased and the survival rate has improved, supporting the optimal follow‐up strategy has become an important issue. This study aimed to evaluate follow‐up imaging usage after breast cancer surgery and the implications on mortality in Korea. Methods This study included 96,575 breast cancer patients diagnosed during 2002–2010 and registered in the Korea Central Cancer Registry, Statistics Korea, and Korean National Health Insurance Service. We evaluated the frequency of breast imaging (mammography and breast MRI) and systemic imaging for evaluating the presence of distant metastasis (chest CT, bone scan, and PET‐CT), and performed analyses to determine if they had an effect on mortality. Results The median follow‐up period was 72.9 months (range: 12.0–133.3) and 7.5% of the patients died. Among all patients, 54.7%, 16.2%, 45.6%, and 8.5% received 3 or more mammograms, chest CTs, bone scans, and PET‐CTs within 3 years after surgery, respectively. Among patients who developed recurrence after 3 or more years, a comparison of overall mortality and breast‐cancer specific mortality according to the frequency of imaging by modality (<3 vs. ≥3) showed that only mammography had significantly reduced mortality (hazard ratio [HR]: 0.72, 95% CI: 0.61–0.84, p < 0.0001; HR: 0.72, 95% CI: 0.61–0.84; p < 0.0001). Conclusions This study showed that only frequent mammography reduced mortality and frequent imaging follow‐up with other modalities did not when compared to less frequent imaging. This finding provides supportive evidence that clinicians need to adhere to the current guidelines for surveillance after breast cancer surgery.


| Study design and population
This BIG-WISE study was approved by the Institutional Review Board of the National Cancer Center, Korea (NCC 2016-0209), and the requirement for informed consent was waived because of the use of de-identified data. The subjects were Korean female breast cancer patients (with International Classification of Diseases, 10th revision [ICD-10] code 12 ) diagnosed between 2002 and 2010 and registered in the KCCR, Statistics Korea, and KNHIS.
Of the 96,575 breast cancer patients initially identified, we excluded 27,031 patients who were male (n = 480), did not undergo breast cancer surgery (n = 12,390), did not have 3 years of data on imaging in the KNHIS (n = 12,393), had distant metastasis at diagnosis (n = 949), or had less than 12 months of follow-up (n = 816) ( Figure 1). Finally, 69,544 breast cancer patients were included in the analysis. Age at diagnosis was classified as <30, 30-39, 40-49, 50-59, 60-69, and ≥70 years. Comorbidities were evaluated using the Charlson Comorbidity Index (CCI) and categorized as 0, 1, and ≥2. Stage at diagnosis, the data on which became available in the KCCR starting in 2005, was classified as local, regional, and missing/unknown following the Surveillance, Epidemiology, and End Results staging system. 13 F I G U R E 1 Study population from National Cohort

| Data sources
For this BIG-WISE study, we established a merged database from three national cohorts: KCCR, KNHIS, and Statistics Korea. KCCR is a population-based national cancer registry that includes information on more than 98% of patients with newly diagnosed cancer in Korea. 14,15 The KNHIS is the single insurer of the Korean public health system, and tracks medical information including cancer treatment status (chemotherapy, radiotherapy, and endocrine therapy) and imaging studies. 16 Mortality data were obtained from Statistics Korea. 17

| Imaging modalities
As follow-up imaging modalities after surgery, breast imaging (mammography and breast MRI), systemic scans to assess the presence of distant metastasis (chest X-ray, chest CT, bone scan, and PET-CT) and treatment-related imaging (DEXA) were evaluated. We investigated the date at which each follow-up modality was performed and the number of studies performed for each modality.
To assess the effect of follow-up imaging on clinical outcomes, patients were divided into two groups according to the number of studies performed for each modality. A cutoff value of 3 was set based on annual check for 3 years. Univariate and multivariate analyses were performed comparing overall and breast cancer-related mortality between the divided groups for each imaging modality. An additional analysis was performed in patients with recurrence after 3 or more years to determine whether the frequent follow-up imaging during 3 years after curative treatment can improve clinical outcomes in recurrent patients.

| Assessment of recurrence and death
Mortality data were obtained from Statistics Korea with the date of the last follow-up as December 31, 2011. 17 The cause of death was recorded and classified according to ICD-10 code. 12 Because KCCR and KNHIS did not have the exact medical information on the date or site of recurrence, we defined recurrence as the presence of new breast cancer surgery, chemotherapy, or radiotherapy, or a change of hormonal therapy, excluding switching and extended therapy, after completion of curative treatments.

| Statistical analysis
Baseline characteristics and imaging modality after surgery were expressed as frequencies in percent. Between-group comparisons of patient characteristics at diagnosis and treatment status (adjuvant chemotherapy, adjuvant radiotherapy, adjuvant hormonal therapy) were performed using the chi-squared test or Fisher's exact test, as appropriate. Cox proportional model was used to evaluate the association between patient survival and the frequency of imaging follow-up for each modality after surgery. Age at diagnosis, treatment status (chemotherapy, radiotherapy, adjuvant hormonal therapy), and CCI without cancer were adjusted for in the multivariable Cox proportional hazard model. All statistical analyses were performed using SAS 9.4 (SAS Institute Inc.), and a p value <0.05 was considered statistically significant. The median follow-up period was 72.9 months (range: 12.0-133.3), and 11,676 patients (16.8%) experienced recurrence and 5241 (7.5%) died, 83.1% of whom died from breast cancer.  Table 3). Patients with local disease underwent more mammography (14,987/25,583, 58.6%) and breast MRI (1281/25,583, 5.0%). Systemic imaging work-ups were performed in patients with regional disease, those who underwent chemotherapy, and those who underwent radiotherapy (

| Survival according to pattern of imaging work-up
Our analysis of the association between imaging followup and clinical outcomes showed that the patients who underwent more systemic imaging follow-up had higher rates of recurrence, breast cancer-related mortality, and overall mortality (Table 4). To determine whether frequent imaging follow-up resulted in a lower mortality rate, we separately compared the HRs for breast cancer-related mortality and overall mortality according to the number of imaging studies performed for each modality (<3 vs. ≥3) within 3 years among patients who developed recurrence 3 years after surgery (Table 5). In univariate analyses, patients who underwent mammography ≥3 times showed a lower overall mortality rate (HR: 0.69, 95% CI: 0.59-0.81, p < 0.001) and breast cancer-related mortality rate (HR: 0.69, 95% CI: 0.58-0.81, p < 0.001) compared to those who underwent mammography <3 times. After adjusting for age at diagnosis, chemotherapy, radiotherapy, adjuvant hormonal therapy, and comorbidities, only frequent mammography significantly influenced overall mortality (HR: 0.72, 95% CI: 0.61-0.84, p < 0.001) and breast cancerrelated mortality (HR: 0.72, 95% CI: 0.61-0.84, p < 0.001). After additionally adjusting for the stage at diagnosis, any breast imaging including mammography and systemic imaging did not influence survivals (Table 5).

| DISCUSSION
This study found that among the imaging modalities used for follow-up surveillance after curative surgery for breast cancer, only frequent mammography is associated with survival, and frequent use of other imaging modalities did not lower the rates of overall and breast cancerrelated mortality, particularly in patients who developed recurrence after surgery. Further, the pattern of imaging follow-up differed by age, stage, and type of treatment.       Recent guidelines, such as those from the American Society of Clinical Oncology, National Comprehensive Cancer Network (NCCN), and European Society for Medical Oncology (ESMO) do not recommend other laboratory tests (e.g., tumor markers) or imaging tests (e.g., bone scans, chest or abdominal CT, PET-CT) in asymptomatic patients because there is no evidence to support their survival benefit. 5,18,19 However, these tests are being performed in clinical practice because of patient and clinician fear of recurrence or metastasis and the belief that early detection using more intensive imaging work-ups reduces cancer-related death. This is supported by the results of the current study, in which 97.2% of the patients underwent a systemic imaging work-up that included a bone scan, CT, or PET-CT. In a survey of clinicians on follow-up after primary treatment of breast cancer conducted by the Korean Breast Cancer Society, most respondents indicated that they conducted more intensive follow-up imaging work-ups than recommended in the current guidelines, 11 similar to the findings of the current study.
With respect to the patterns of imaging work-ups, we found that they differed by age, stage, and type of treatment. In young women aged <30, more breast MRIs and fewer mammograms were performed. This could be because of the tendency for dense breast tissue and genetic susceptibility (BRCA1/2 mutation) in these patients. 20,21 Meanwhile, in patients treated with adjuvant hormonal therapy (i.e., those with hormone receptor-positive breast cancer), more bone scans and fewer chest CTs were performed. This result could be interpreted to mean that clinicians tend to recommend different systemic imaging modalities according to the tumor characteristics, as hormone receptor-positive tumors develop more bone metastases than visceral metastases.
We also found that systemic imaging work-ups were more frequently performed in patients with advanced cancer and in patients who receive chemotherapy. However, this did not improve overall survival or breast cancerrelated survival. Previous randomized controlled trials have found that less-intensive follow-up strategies did not negatively affect patient outcomes or early detection of recurrence. In addition, more intensive follow-up was associated with higher costs without differences in early detection of relapses. 22,23 In other systematic review article evaluating the clinical effects of intensive versus less-intensive follow-up on disease outcomes, intensive follow-up with more frequent work-ups did not reduce mortality and recurrences in breast cancer patients. 24 In addition, there was no survival benefit associated with the early diagnosis of recurrence by intensive follow-up prior to the occurrence of symptoms, supporting the validity of the current guidelines. 9 Moreover, one study reported that the intensity of imaging work-up did not affect QoL  in breast cancer survivors, and 70% of the patients even reported feeling more stressed and anxious when they visited clinics, especially after undergoing tests. 25 Expert panel on breast imaging according to ACR appropriateness criteria comment that there is no role for imaging to screen for distant recurrences in asymptomatic patients with a history of stage I breast cancer that received treatment for curative intent. They recommend that routine surveillance with an annual mammogram is the only imaging test that should be performed to detect an in-breast recurrence or a new primary breast cancer in women with a history of stage I breast cancer. 26 This study has some limitations. First, the merged data used for the study did not include detailed information on tumor characteristics, such as hormone receptor status. To compensate, we analyzed the outcomes by adjusting for hormonal therapy status. Second, we were not able to take into account the indication for the imaging work-ups such as cancer-related symptoms, comorbidity-related causes, or other medical issues or who prescribed the work-ups (e.g., oncologists or primary physician) in our analyses. However, we considered the presence of morbid disease and the cause of death in our analysis of the effect of intensive imaging work-ups on survival. Third, this study was not randomized or a planned prospective study. However, the national health insurance data and statistics which our analyses are based on are highly specific because the KNHIS is the only public health insurer and covers at least 98% of Koreans and is the only representative national database to include cause of death. 27 In conclusion, this study showed that as a follow-up imaging modality, only frequent mammography and no other imaging modalities reduce overall mortality and breast cancer-related mortality in Korean female breast cancer patients. These findings provide evidence that frequent systemic imaging work-ups are not needed despite the fear of recurrence. Therefore, clinicians need to adhere to the current guidelines for surveillance after curative treatment in breast cancer patients.