ABSTRACT

Objective:

To assess the impact of the coronavirus disease-2019 (COVID-19) pandemic screening restrictions on the diagnosis and treatment of breast cancer in a single health system.

Conclusion:

Despite an increase in tumor size and positive nodal status seen during the screening moratorium, surgical treatment was not negatively impacted.

Results:

Fewer patients presented with a breast cancer diagnosis during Early phase 2020 when compared to any other time period. Numbers increased significantly in Late phase 2020; total numbers of patients seen in 2020 approached but did not completely reach that of 2019. When compared to other time periods, patients who presented during the moratorium on screening were younger, more likely to be black, had a higher Body Mass Index, and were more likely to have a human epidermal growth factor receptor 2 positive tumor. There was a slight increase in size of presenting tumor and node positivity, although no differences in breast or axillary surgical management were identified.

Materials and Methods:

We performed a retrospective, cohort investigation of breast cancer patients at a multi-institution health system from March 1, 2019 to December 31, 2020 with two time periods related to the pandemic: “Early phase” (March 18 – June 7) reflecting the time of the screening mammography moratorium and “Late phase” (June 8 – December 31) to reflect the time once screening mammography resumed. 2020 was compared to 2019 to exclude potential differences from temporal or seasonal changes. Variables included demographics, COVID related-deferral, cancer specific data, method of detection, type of treatment recommended and received.

Key Points

• A slight increase in presenting tumor size and positive nodal status was identified after screening mammography was halted.

• Although endocrine therapy was offered as a temporizing measure during the lockdown, there was low uptake.

• Ultimate surgical treatment was not impacted by screening cessation.

Introduction

Coronavirus disease-2019 (COVID-19) has affected healthcare delivery more than any other crisis in recent memory. The World Health Organization first announced concerns about a coronavirus-related pneumonia in Wuhan, China on January 5, 2020. The first documented COVID-19 positive patient in the United States occurred on January 21, 2020 (1). In Massachusetts, the first case was documented on February 1, 2020 (2). A state of emergency was declared on March 10, 2020 with cessation of all elective procedures as of March 16, 2020. Screening imaging was halted at our institution on March 18, 2020 and resumed June 8, 2020.

Routine screening mammography has resulted in earlier detection of breast cancer and a reduction in the extent of treatment. Screening has been so effective that the benefit of self breast exam (SBE) and clinical breast exam (CBE) have been called into question (3, 4). The American Cancer Society currently advises against CBE in women undergoing screening and against SBE for women of any age (5). Some studies have questioned the benefit of SBE even in regions of the world where mammography is not readily available, as discussed in a 2003 Cochrane review (6).

Screening imaging cessation would be expected to have an impact on the detection of early breast cancer and therefore result in a more advanced stage at presentation and worse outcomes (7, 8, 9, 10). What is less clear is how COVID-19 restrictions would affect the number of patients presenting with palpable (and likely more advanced) cancers in a population accustomed to screening mammography and discouraged to perform SBE.

The goal of this study was to assess the impact these restrictions had on the diagnosis of breast cancer presentation and the therapies offered. We hypothesized that the disruption of the availability of screening imaging and “routine” CBE would result in a delay in the diagnosis of breast cancer, increased stage at presentation, and altered treatment recommendations. We also hypothesized that there would be a decrease in patients presenting with breast cancer, due both to a lack of screening and patients purposefully delaying the evaluation of palpable abnormalities due to fear of contracting COVID-19 while seeking medical care.

Materials and Methods

We conducted an Institutional Review Board (IRB)-approved retrospective, cohort study at a 720-bed tertiary care center with three regional hospitals in Western Massachusetts. We identified patients presenting with a new breast cancer diagnosis from March 1, 2019 to December 31, 2020 using the institution’s IRB-approved Breast Disease Patient Repository, a secure, HIPAA compliant REDCap database, which is prospectively maintained. All patients with a new breast cancer diagnosis who presented between March 18 – December 31, 2020 were included in the study and compared to all patients with a breast cancer diagnoses who presented between March 18 – December 31, 2019. March 18^th, the first day of the screening moratorium in 2020, was chosen as the start date. Exclusion criteria were those patients with breast cancer who presented outside this time frame. Supplemental information was obtained from the health system’s electronic medical record.

We created two time period groups referred to as “Phases”. The first time period reflected the pause in screening mammography (Early phase: March 18 – June 7, 2020) and the second time period reflected screening mammography resumption (Late phase: June 8 – December 31, 2020). We compared groups from 2020 (during COVID) to 2019 (before COVID) to assess whether any potential differences were due to COVID-19 and not to temporal or seasonal changes.

Eligible encounters were uploaded to a REDCap database, hosted by Tufts Clinical and Translational Science Institute (Grant Number UL1TR001064) for abstraction from the electronic medical record. Variables collected included patient age, gender, ethnicity, Body Mass Index (BMI), COVID deferral (treatment was treatment delayed or not), cancer specific data [specifically AJCC 8^th edition clinical stage, grade, hormone receptor status and human epidermal growth factor receptor 2 (HER2) status], and method of detection (including imaging, self-detected, clinically detected). The type of treatment (surgery first versus neoadjuvant therapy) and the type of neoadjuvant therapy (chemotherapy versus endocrine therapy) that would have been recommended if the COVID pandemic had not occurred, as well as the surgical treatment of breast (lumpectomy, mastectomy, none) and surgical treatment of axilla (sentinel node biopsy, axillary dissection, completion axillary dissection, none) that patients ultimately received were also collected.

Results

We identified a total of 583 patients with breast cancer who presented between March 18, 2019 and December 31, 2020. In 2019, Early phase and Late phase consisted of 88 and 217 patients, respectively, for a total of 305 patients, whereas in 2020 Early and Late phase included 27 and 252 patients, respectively, for a total of 279 patients. Demographics, clinical characteristics, cancer specific data and treatment data for the study population are shown in Table 1. Patients who presented during Early phase 2020 were younger (p<0.01) and were more likely to be black (p = 0.05) than during the other three phases. Tumors were more likely to be HER2 positive (p<0.01) as seen in Figure 1. In the cohort analysis, there was no difference in tumor size (p = 0.24) or lymph node positivity (p = 0.11). Metastatic disease at presentation was equally infrequent among all phases. There was no difference in the type of breast surgery (p = 0.95) or axillary treatment (p = 0.39) that the patients ultimately received, regardless of the pandemic, as seen in Figure 2.

Sensitivity analysis was performed to compare the period of the moratorium on screening mammography (Early phase 2020) against the other phases (Tables 2, 3, 4). This confirmed the absence of a treatment difference in the surgical management of the breast and axilla that was seen in analysis of the entire cohort (Table 1), even when other differences were noted.

During the initial screening moratorium (Early phase 2020), only 27 patients presented with breast cancer compared to 88 patients in the same time period in 2019 (Table 2). Patients in Early phase 2020 who presented with a new diagnosis of breast cancer were noted to have a higher BMI (30.6 versus 28.1, p = 0.05).

All tumors were self-detected (100%) during Early phase 2020, compared with 28% (n = 25) in the same time period the year prior. There was no difference in tumor type, grade or receptor status when compared to the tumors diagnosed the previous year, but a slightly larger tumor size was observed among the new diagnoses during the pandemic (average T-stage of 1.5 versus 1.1, p = 0.05, Table 2 and Figure 3) although analysis of the entire cohort included in the study did not show significance (p = 0.24, Table 1).

During Early phase 2020, patients were more likely to be treated with neoadjuvant therapy (37.0 versus 17.0%, p = 0.03). There was no difference between the type of neoadjuvant treatment that was chosen (endocrine versus chemotherapy, p = 0.67). Only 6 out of 21 patients (21.4%) who were ER positive agreed to take endocrine therapy during the deferral period. There was no difference in surgical treatment of the breast or the axilla when surgery was eventually performed (Table 2).

Screening mammography resumed in the latter part of 2020 and was compared to the same time period the previous year (Late phase 2019 to 2020, Table 3). Patients diagnosed with breast cancer in the second half of 2020 were older than those diagnosed the year before (64 versus 61.5 years, p = 0.04). All other clinical characteristics, T- and N- stage and treatment types were similar.

When Early phase 2020 was compared to Late phase 2020 (Table 4), patients who presented with breast cancer in the first part of the year were younger (54.8 versus 64 years, p<0.01). HER2 positivity was higher in Early phase 2020 (11.5 versus 8.2%, p = 0.02) but there were no differences in grade or hormone receptor status between groups. The average N stage was higher when compared to after resumption of screening (0.2 versus 0.1, p = 0.02) but this did not increase use of axillary node dissection (p = 0.39). Neoadjuvant therapy was recommended in 37.0% of cases during the Early phase 2020 compared with 17.0% during the Late phase 2020 (p = 0.01).

Discussion and Conclusion

The abrupt cessation of screening imaging and elective procedures immediately caused concern about worse cancer outcomes. Within one week of the lockdown, a strategy was developed locally to address management of new cancer patients by optimizing use of endocrine therapy where possible until the COVID-19 Pandemic Breast Cancer Consortium recommendations were released (11). Other guidelines were generated over the next several months that supported similar strategies (12, 13, 14).

During the Early phase 2020, which represents our time of strict COVID lockdown, we observed a decrease in the number of patients with a new breast cancer diagnosis. As expected, all newly diagnosed cancers were self-detected with a statistically significant difference in mean tumor size. It is not surprising that more of these self-palpated cancers were HER2 positive which is indicative of a more aggressive subtype. We hypothesize that patients, upon learning about the pause of routine screening, were more likely than before to perform a self-examination and, as a result, were noticing these tumors. Once routine screening resumed, however, no significant difference in tumor size was seen between patients whose cancers were detected by imaging compared to self-examination.

Neoadjuvant therapy was recommended more often in the first phase, as surgeries including for oncologic reasons were halted at this time. The impact of delay to surgical treatment in breast cancer patients has been studied in large datasets and is predicted to result in worse all-cause mortality (15, 16). The effects of the pandemic-related treatment delays on survival and recurrence in patients who chose to decline neoadjuvant endocrine therapy, as was our experience, will have to be further evaluated in the future. In Late phase 2020, it was determined that breast surgery could occur safely (17, 18). Oncologic teams resumed pre-pandemic protocols which allowed most patients to have surgery as a first treatment.

Similar to our local experience, screening mammography rates nationally increased and remain elevated once moratoriums were removed despite the persistence of COVID, although underserved populations are less likely to resume screening or more likely to cancel and not reschedule (19, 20, 21). We planned for the resumption of screening almost as soon as the moratorium started, due to recognition of the importance of planning and messaging (22, 23). Nearly the same number of mammograms were performed in 2020 as in 2019 using extended hours and weekend schedules to accommodate social distancing guidelines and the backlog patients. Our institution diagnosed 279 patients with breast cancer during the pandemic in 2020, 92% of the number of patients diagnosed in 2019 and less of a decrease than we had feared.

Once screening imaging or routine clinical examination is not available, patients become reliant on self-examination for cancer detection. During the Early phase 2020, the time of strict COVID lockdown, the number of patients presenting with a new cancer diagnosis decreased as all newly diagnosed cancers were self-detected. A small, statistically significant difference was seen in mean tumor size, but this did not impact ultimate surgical treatment. No significant difference in tumor size was seen between patients whose cancers were detected by imaging compared to self-examination once imaging was performed. In Late phase 2020, care returned to normal and most patients underwent surgery first when it was found that breast surgery could occur safely (17, 18).

Tonneson et al. (24) did not see a difference in the stage of presentation when looking at patients who presented between March and August 2020. We were able to also look at patients in the six months after screening resumed to determine if there was any difference in presentation. Sensitivity analysis revealed a marginal difference in T-stage (1.5 versus 1.1, p = 0.08) and a small but statistically significant difference in N stage (0.2 versus 0.1, p = 0.02). Ultimately, we did not see a resulting difference between lumpectomy and mastectomy rates, nor was there a difference in axillary treatment (Figure 2).

Early-stage breast cancer diagnosis relies on effective screening programs, facilitates greater rates of breast conservation and allows some women to avoid radiation and axillary sentinel node biopsy as part of the Choosing Wisely campaign (25-27). A logical consequence of delayed screening during the pandemic would be a subsequent increase in mastectomy rates in the setting of higher numbers of palpable advanced stage cancers, for which breast conservation is not an option. Although other studies have shown an impact on surgery (28, 29), we did not observe that impact.

It should be noted, however, that the findings of this study do not negate or refute the established data on the mortality reduction seen with established screening mammography protocols. Our data reflects short-term mammography cessation of three months and supports the need for quick resumption of screening to prevent longer delays in cancer detection. Even a short period of screening stoppage can result in a longer delay to presentation due to patient hesitance. Studies that look at longer interruptions in screening showed significantly worse alterations in both stage and surgical treatment (30). Concerns about the need for a proactive approach by radiology to ensure timely screening resumption are well documented (31).

This study reflects a single institution’s experience with breast cancer and the moratorium against screening imaging during the initial phases of the COVID-19 pandemic. As a result, the single institution design limits generalizability and small numbers preclude discrete statistical analysis.

Despite these limitations, our study adds important information and raises points for discussion. It may be appropriate to revisit the recommendation to avoid self-examination, as it can be a very valuable tool to detect new breast cancer, especially when routine screening is not available. Our study contributes valuable data to evaluate the impact of short interruptions to breast cancer radiology screening on stage at diagnosis. We did not see any difference between lumpectomy and mastectomy rates or axillary management when comparing Early or Late phase by years or by yearly totals themselves. At the time of manuscript preparation, there was no published literature specifically evaluating the impact of the pandemic on breast conservation rates or axillary management with such a short interval of screening deferment; thus, this paper is additive to available information. The implications of these findings are still unclear. Distinct differences in cancer presentation during the initial pandemic phase were observed, but these did not appear to be associated with clinically significant differences in treatment. Additional long-term follow-up is necessary to determine the impact of this screening moratorium and the resulting treatment delays on breast cancer recurrence and survival.

In conclusion, patients who presented with breast cancer during the COVID-19 pandemic in the absence of screening mammography were more likely to be younger, have a higher BMI, present with HER2 positive cancers, be node positive, and receive neoadjuvant treatment most commonly with endocrine therapy. Despite these differences, ultimate surgical management was not impacted by pandemic-related screening cessation.

Statistical Analysis

All variables were checked for completeness and plausibility using frequencies (percentage, categorical) and means/ranges (continuous, ordinal). Descriptive statistics were calculated for baseline time periods from 2019 (prior to the COVID pandemic) and 2020 (after the onset of the COVID pandemic), including percentages for binary categorical variables, means (standard deviation) and medians (interquartile range) for continuous variables. The t-test was used for continuous variables and Fisher’s Exact test for categorical variables. In order to further evaluate temporal trends within our data, we conducted stratified analysis among Early phase 2020 versus Early phase 2019, Late phase 2020 versus Early phase 2019 and Early phase 2020 versus Late phase 2020. Statistical significance was set at an alpha of 0.05. Data were analyzed using STATA 16 (StataCorp, College Station, TX, USA).