Treatment times in breast cancer patients receiving neoadjuvant vs adjuvant chemotherapy: Is efficiency a benefit of preoperative chemotherapy?

Abstract Background/Objective Delays in times to surgery, chemotherapy, and radiotherapy impair survival in breast cancer patients. Neoadjuvant chemotherapy (NAC) confers equivalent survival to adjuvant chemotherapy (AC), but it remains unknown which approach facilitates faster initiation and completion of treatment. Methods Women ≥18 years old with nonrecurrent, noninflammatory, clinical stage I‐III breast cancer diagnosed between 2004 and 2015 who underwent both surgery and chemotherapy were reviewed from the National Cancer Database. Results Among 155 606 women overall, 28 241 patients received NAC and 127 365 patients received AC. NAC patients had higher clinical T and N stages (35.8% T3/4 vs 4.9% T3/4; 14.4% N2/3 vs 3.7% N2/3). After adjusting for stage and other factors, NAC patients had longer times to begin treatment (36.1 vs 35.4 days adjusted, P = .15), and took significantly longer to start radiotherapy (240.8 vs 218.2 days adjusted, P < .0001), and endocrine therapy (301.6 vs 275.7 days adjusted, P < .0001). Unplanned readmissions (1.2% vs 1.7%), 30‐day mortality (0.04% vs 0.01%), and 90‐day mortality (0.30% vs 0.08%) were all low and clinically insignificant between NAC and AC. Conclusion Compared to patients receiving AC, those receiving NAC do not start treatment sooner. In addition, patients receiving NAC do not complete treatment faster. Although there are clear indications for administering NAC vs AC, rapidity of treatment should not be considered a benefit of giving chemotherapy preoperatively.


| INTRODUCTION
Multiple prospective randomized trials 1-3 have demonstrated no differences in overall survival (OS) or improved disease free survival (DFS) in patients receiving neoadjuvant chemotherapy (NAC) vs adjuvant chemotherapy (AC) for the treatment of breast cancer. NAC is being used more commonly, especially in cases of triple negative disease and human epidermal growth factor receptor 2 (HER2) positive disease 4,5 due to the dramatic clinical and pathologic responses often seen. Pathologic complete response rates for triple negative disease range from 23.2% to 33.6%, and for HER2 positive, hormone-receptor negative disease from 38.7% to 66.2%. 6,7 NAC is also appropriate for many breast cancer patients with large primary tumors who desire breast conservation, 1,8,9 and its use may also downstage the axilla before nodal evaluation. 1,8,10 It can also be used with the goal of eliminating systemic micrometastatic disease, 11 and to safely delay surgery in certain situations; for example, by allowing time to medically optimize a patient prior to surgery or providing an opportunity to have genetic testing performed. 9 Additionally, times to treatment have become important, and longer times to surgery, chemotherapy, and radiotherapy all confer modest but significant impairments in survival. 12 While we were unable to find evidence in the published medical literature, it is thought by some that NAC is typically initiated more quickly than proceeding directly to surgery. Despite this view, there is no data available, to our knowledge, regarding the time it takes patients undergoing NAC to start chemotherapy, as vs the time it takes patients to have surgery first. This study was therefore performed to evaluate the times to start and complete breast cancer treatment in each of these settings to assess whether rapidity of treatment is a benefit of NAC.

| MATERIALS AND METHODS
Data were obtained from the National Cancer Database (NCDB). Women ≥18 years old with newly diagnosed, nonrecurrent, noninflammatory, clinical stage I-III breast cancer diagnosed between 2004 and 2015 whose treatment included both surgery and chemotherapy were selected. Patients with recurrent disease, multiple tumors and/or inflammatory breast cancer were also excluded due to their bad prognosis and varied treatment paradigms, and patients in whom a biopsy was not documented in the dataset were excluded. Remaining exclusions are elaborated in Figure 1.
Patients were then divided into two groups; those who received NAC and those who received AC. They were then further stratified based on whether they also received radiotherapy and endocrine therapy ( Figure 2). Treatment times were measured from time of biopsy to start of first treatment (with first treatment being chemotherapy for NAC group, and surgery for AC group). Time from biopsy to the start of radiotherapy was also measured for all patients who received radiation. Time from biopsy to the start of endocrine therapy was also analyzed. The start of endocrine therapy was used as a surrogate for the end of treatment. Thus, only patients who received endocrine therapy (n = 52 264) were included in that portion of the analysis.
Comparisons between the groups were performed using Student's t test for continuous variables and chi-squared test for categorical variables. To minimize the potential selection bias, treatment times were adjusted for facility volume, age, race, education, insurance, income, setting (urban vs rural vs metropolitan), facility distance, treatment at more than one facility, Charlson comorbidity index, histology, F I G U R E 1 Exclusion diagram grade, clinical T stage, clinical N stage, AJCC clinical stage, pathologic T stage, pathologic N stage, and phenotype (estrogen, progesterone and HER2) by propensity score with 1:1 matching method. Multiple regression models with log-transformed treatment times (due to skewed distribution) and with length of hospital stay (approximately normal distribution), as well as multivariable logistic regression models with readmission and 30-and 90-day mortality were constructed to determine predictors associated with these clinical outcomes.
All analyses were performed using SAS 9.4 (SAS Inc). Statistical tests were two-sided, and a P-value of <.05 was considered statistically significant.

| Times to treatment
Unadjusted time comparisons were first determined ( Table  2). Time from biopsy to first treatment was 35.6 ± 27.5 days in patients treated with NAC, vs 33.4 ± 22.9 days in patients treated with AC (P < .0001). Unadjusted time from biopsy to radiation was 243.2 ± 58.8 days in NAC vs 208.7 ± 54.6 days in AC (P < .0001), and unadjusted time from biopsy to start of endocrine therapy was 305.4 ± 77.6 days in NAC vs 268.3 ± 71.1 days in AC (P < .0001). After propensity score matching, adjusted times are detailed in Table 2. NAC patients had a similar time to begin treatment compared to AC patients, but took significantly longer to start radiotherapy and endocrine therapy.

| Factors that influence treatment times
As depicted in Figure 3, predictors of times to treatment initiation were examined. The factors associated with a longer time to treatment initiation included higher volume facilities, black and Hispanic patients, increased income, and increased Charlson comorbidity index. Patients treated at more than one facility, along with black and Hispanic patients were more likely to have longer times to treatment initiation as well as delays in starting radiation and endocrine therapy. Factors associated with shorter times to first treatment, radiation therapy, and endocrine therapy included higher education level, private insurance, and treatment in an urban or rural setting (as opposed to a metropolitan setting).

| Outcomes
After propensity score matching, there was no difference in the length of hospital stay among patients treated with NAC vs AC (1.16 ± 4.93 vs 1.28 ± 5.90 days, P = .14). Rates of unplanned readmission were slightly lower in patients treated with NAC vs AC (1.18% vs 1.72%), however this was not clinically significant. Similarly, 30-day mortality rates (0.04% vs 0.01%) and 90-day mortality rates (0.30% vs 0.08%) were also not clinically different since the rates were so low.

| Independent predictors of increased length of hospital stay, readmission & mortality
Multivariable analyses were performed to determine independent predictors of increased length of hospital stay, readmission, and 30-and 90-day mortality rates. Predictors of increased length of hospital stay were Charlson comorbidity score of 3 (0.36 days longer, 95% confidence interval [CI] 0.01-1.19, P = .04), F I G U R E 2 Characterization of patients based on treatment sequence. Patients were initially divided into 2 categories, based on whether they were treated with neoadjuvant chemotherapy or adjuvant chemotherapy. Within those groups, they were then further characterized into 3 groups: Those treated with only surgery and chemotherapy, those who also had radiation, and those who also had radiation and endocrine therapy. These individual groups represent the cohorts included in this study. Times to start treatment (time from biopsy to surgery, or from biopsy to chemotherapy) included all groups. Times to radiation included only those having that modality. Times to complete treatment were measured in those having endocrine therapy. Abbreviations: AC, Adjuvant chemotheraphy; B, Time of biopsy; C, Chemotherapy; E, Endocrine theraphy; NAC, Neoadjuvant chemotheraphy; R, Radiotheraphy; S, Surgery.

| DISCUSSION
NAC has benefits in the management of breast cancer including shrinking the size of the primary tumor to allow for increased rates of breast conservation, and downstaging disease in the axilla. Some proponents of NAC also believe that it shortens the time to treatment initiation compared to upfront surgery, however, to our knowledge, there is no data substantiating this claim, even though the published medical literature does provide data relating to general delays in breast cancer care. 13,14 The interest in whether NAC improves time to treatment initiation likely originates from data showing that a longer time to surgery is associated with lower OS and DFS in breast cancer patients, with the largest decline seen in patients having stage I and II disease. 13 While delays to surgery are significant, a delay of >90 days as vs ≤30 days has been found to be associated with only a 3%-5% decrement in survival. 13 Similarly, a delay of more than 90 days

T A B L E 2 Unmatched and propensity
score-matched time comparisons to initiation of chemotherapy is also associated with worse OS and breast cancer-specific survival, 15,16 and a study evaluating patients with hormone receptor-negative breast cancer found that delaying chemotherapy by 6 weeks resulted in worse OS. 17 Regarding delays to initiation of adjuvant radiotherapy, the data is mixed, with some studies suggesting that outcomes are unaffected when radiation is given between 8 and 20 weeks postoperatively in patients who do not receive chemotherapy, and within 365 days in patients who do receive chemotherapy. 12 For these reasons, quality measures currently specify three recommended time intervals 18 : firstly, that radiotherapy should be given within 365 days of diagnosis in women under the age of 70 who have breast conserving surgery. Secondly, that chemotherapy should be considered or administered within 4 months of diagnosis for women under the age of 70 with AJCC T1c, Stage II, or Stage III hormone-receptor-negative breast cancer, and thirdly, that tamoxifen or a third-generation aromatase inhibitor is considered or administered within 1 year of diagnosis for women with AJCC T1c, Stage II, or Stage III hormone receptor positive breast cancer. Compliance with these measures is tracked by accrediting bodies such as the National Accreditation Program for Breast Centers 18 because of the importance of appropriately initiating treatment without undue delay. Avoiding treatment delays is therefore important to both optimize outcomes and comply with quality measures.
After adjusting for numerous factors, we found that time from biopsy to treatment initiation (either surgery or chemotherapy) was similar in patients receiving NAC vs AC. Thus, NAC does not expedite time to treatment initiation, even though there may be situations where NAC may be desirable if undue preoperative delays are unavoidable.
Delays to surgery also have an impact on time to chemotherapy and this could affect compliance with the chemotherapy quality measure. Patients who undergo mastectomy with or without reconstruction are at a higher risk of having delays to AC due to longer recovery time and higher rate of postoperative complications. 19 For patients undergoing larger procedures where postoperative recovery time is predicted to be unusually prolonged, it may be reasonable to treat them with chemotherapy upfront. It must be remembered that times to treatment are likely eclipsed by the "silent interval" of a tumor, which is defined as the period of tumor growth beginning at inception, up until it can be detected. By definition, we cannot know the length of this period of time, but it is generally thought that these intervals frequently comprise the majority of a tumor's lifespan. 20 It is because of this that delays to surgery, chemotherapy or radiotherapy likely have much less of an impact than typically surmised.

T A B L E 3 Independent predictors of increased length of stay, readmission & mortality
We found that the time to initiation of radiotherapy was 3.2 weeks longer in the NAC group than in the AC group, however, this delay of 3 weeks in the NAC cohort is unlikely to impact clinical outcomes in most cases. Most published studies have found that delays in radiotherapy of at least 8-20 weeks are required before a decline in survival or an increase in local recurrence is seen. 12 However, a study evaluating 581 patients specifically treated with NAC and postoperative radiotherapy found that initiating radiation <8 weeks after surgery was associated with improved disease-specific and OS. 21 Ultimately, with multidisciplinary treatment, the impact of longer radiotherapy after changing the order of surgery and chemotherapy remains uncertain. However, the current radiotherapy standard requires that radiation be administered within 1 year of diagnosis, 18 and so clinicians must, at minimum be cognizant of the fact that a 3-week delay can mean the difference between compliance and noncompliance in terms of quality measures.
A second objective of our study was to assess the time to completion of treatment, and the start of endocrine therapy was used as a surrogate for this time point. Since endocrine therapy is given as an outpatient medication, compliance is difficult to monitor, and lengths of therapy can vary from 5 to 10 years, the start, rather than the completion, of hormonal therapy was felt to be a reliable estimate of the completion of their overall treatment course. Moreover, we are not aware of any similarly sized dataset that assesses either time to completion of endocrine therapy or the level of compliance. We found that time to completion of treatment was 3.7 weeks longer in the NAC group compared to the AC group. Unfortunately, there is no good data, to our knowledge, regarding delays in endocrine therapy on outcomes. Especially with time spans for hormonal therapy on treatment being 5-10 years, delays over a matter of weeks likely make little difference. Moreover it must be remembered that the expansion of endocrine therapy from 5 to 10 years was to solely provide a 3% survival benefit, 22 making any delays seen here likely inconsequential. We therefore do not suggest that the time delay to endocrine therapy, which is our surrogate for completion of treatment, in the NAC group has impaired survival, but instead simply that these data demonstrate that it has not expedited or improved outcomes related to potential delays.
In an effort to provide a potential explanation for differences between the NAC group and those having surgery first, we speculated that NAC might increase the risk of perioperative complications, resulting in increased length of hospital stay after surgery due to wound complications or other chemotherapy-related side effects. However our analysis showed no difference in length of hospital stay between AC and NAC. This is consistent with prior studies that have found that NAC is not associated with an increase in short-term complications after mastectomy and implant-based reconstruction 23 or autologous tissue reconstruction. 24 Similarly, we also examined whether the longer time to complete treatment in the NAC group was due to higher readmission rates, but surprisingly, patients treated with NAC actually had lower rates of both unplanned and planned readmission. The reason for this is unclear, but perhaps this could be due to the closer follow up and monitoring of these patients while they were undergoing chemotherapy. With regard to mortality rates, as expected, 30-and 90-day mortality rates were very low and clinically insignificant. We therefore cannot attribute the longer time to completion of treatment to an increase in perioperative complications resulting from NAC. The small differences in timing are therefore more likely due to scheduling factors or even due to a self-imposed delay to wait for recovery from chemotherapy effects before embarking upon surgery out of concern for wound healing.
We found that factors having the greatest correlation with a longer time to treatment initiation were higher volume facilities, increased income, and increased Charlson comorbidity index. Higher volume facilities may have longer wait times for consultations, resulting in delays to surgery or starting NAC. However, treatment at high volume centers has been shown to be associated with improved OS. 25 It is unclear why higher income was associated with longer treatment times, but the most likely reason is that these patients have the financial means to obtain multiple opinions at different facilities, which lengthens the time from diagnosis to treatment initiation. 26 Lastly, increased Charlson comorbidity index was also associated with a longer time to treatment initiation, presumably because these patients require more extensive evaluation and clearance before starting NAC or before proceeding to surgery.
We also found that patients treated at more than one facility, along with black and Hispanic patients had longer times to first treatment, radiation therapy and endocrine therapy, consistent with known disparities and the published medical literature. Stage for stage, black women have been found to have higher breast cancer mortality rates than white women, 27,28 which could be a result of either differences in access to care, or disparities in how care is delivered. Studies have also demonstrated that time from biopsy to treatment initiation is longer in black women 27 and blacks and Hispanics have been found to have greater delays to surgery, 14 all consistent with our findings here.
In contrast to income level, higher education level and having private insurance were associated with a shorter time to treatment initiation, radiotherapy, and endocrine therapy which might be explained, respectively, by knowledge of the disease and the imperative to be treated, and having access to more specialists to facilitate care. Being treated in an urban or rural setting as opposed to a metropolitan setting also resulted in shorter treatment times. Despite the smaller volume and potential shorter wait to see a physician, there are likely certain geographical areas which are lacking breast cancer specialists and have a longer delay. All of these findings indicate that, unfortunately, disparities remain in the neoadjuvant setting as well, and are not dependent on or related to treatment order.
Although this study utilizes a large prospectively collected validated dataset, it is limited by the NCDB's accuracy, which is dictated by how well the cases are coded into it. Additionally, like all datasets, the NCDB dataset is limited to specific variables, and unknown confounders could exist. Length of chemotherapy treatment, type of chemotherapy administered, and time to completion of endocrine therapy are not collected in this dataset, thus limiting our ability to assess how often truncation of those therapies occurs, and how timing is consequently affected. To overcome this, the time to initiation of endocrine therapy was used as a surrogate for the end of treatment. Our study's strengths include the very large size of the dataset utilized and its applicability to the national population at large, as well as its ability to discern precise intervals of treatment.

| CONCLUSION
While there are clear indications for NAC in the treatment of breast cancer, NAC did not result in patients starting or completing treatment faster than those who received AC. Although one might consider its use if prolonged preoperative or postoperative delays are expected, quicker treatment initiation and completion should not be considered a routine benefit of NAC as vs primary surgery, and so this is not an indication to administer chemotherapy upfront.