Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2026 Jul 3.
Published in final edited form as: Cancer. 2026 Mar 15;132(6):e70318. doi: 10.1002/cncr.70318

The landscape of chemotherapy modifications among women treated for stage I-IIIA breast cancer

Jenna Bhimani 1, Kelli O’Connell 1, Victoria S Blinder 2, Rachael Burganowski 3, Isaac J Ergas 4, Grace B Gallagher 1, Jennifer J Griggs 5, Narre Heon 1, Sankeerth Jinna 1, Tatjana Kolevska 6, Yuriy Kotsurovskyy 1, Candyce H Kroenke 4,7, Cecile A Laurent 4, Raymond Liu 4,8, Kanichi G Nakata 3, Sonia Persaud 1, Janise M Roh 4, Emily Valice 4, Elisa V Bandera 9, Erin J Aiello Bowles 3, Lawrence H Kushi 4, Elizabeth D Kantor 1
PMCID: PMC13326849  NIHMSID: NIHMS2149060  PMID: 41808563

Abstract

Introduction

In the delivery of cancer care, patients may experience modifications to their intended chemotherapy regimen for reasons such as toxicity, tolerability, and scheduling challenges. Understanding modifications from intended chemotherapy in the real-world setting of cancer care may affect treatment planning and inferences of real-world treatment effectiveness.

Methods

Leveraging data from the Optimal Breast Cancer Chemotherapy Dosing (OBCD) study, we studied 12,558 women receiving chemotherapy for primary stage I-IIIA breast cancer at Kaiser Permanente Northern California (2006–2019) and Kaiser Permanente Washington (2004–2015). We described modifications in chemotherapy administered from intended, both overall and by regimen. Prevalence ratios (PR) were calculated for associations between regimen attributes and treatment modifications using generalized linear models of the Poisson family with a log link-function and robust standard errors.

Results

Intent to receive a guideline-concordant regimen was associated with a lower likelihood of modifications from intended chemotherapy (vs non-guideline-concordant) (PR:0.74; 95% CI:0.69–0.79), as was receiving a dose-dense regimen (vs non-dose dense) (PR:0.84; 95% CI:0.78–0.91). Those receiving any neoadjuvant chemotherapy (vs adjuvant only) were more likely to have modifications from intended chemotherapy (PR:1.29; 95% CI:1.19–1.41). Increasing number of cycles and drugs administered were associated with a higher likelihood of modifications (p-trend<0.001).

Conclusion

Breast cancer chemotherapy regimens used in the real-world setting were examined. Chemotherapy modifications varied markedly by regimen and characteristics of regimen administered. Given that more than one in three women in this cohort did not receive chemotherapy as intended, these novel findings may be informative in treatment planning, and for future research assessing real-world treatment effectiveness.

Keywords: Breast Cancer, Breast Neoplasm, Chemotherapy, Chemotherapy modification, Electronic Health Records, Treatment

Précis (condensed abstract):

Chemotherapy modifications varied markedly by regimen and characteristics of regimen administered. Given that over 1 in 3 women did not receive chemotherapy as intended, these novel findings may be informative in treatment planning, and to research assessing real-world treatment effectiveness.

Introduction

The evidence base for breast cancer treatment is driven by randomized control trials (RCTs). Thus, extrapolating outcomes from trials to real-world populations requires aligning treatment administration with trial protocols.1 However, trials are often conducted in populations that are healthier and younger than real-world populations, with disparities in enrollment among minoritized groups.2,3 Although modifications from planned care can occur in trials, under-represented patient groups (including older and comorbid adults) experience more toxicity and challenges with tolerability.4,5 Additionally, real-world patients may have more scheduling conflicts driven by employment and caring responsibilities compared to trial participants, ultimately impacting dose reductions, delays, regimen changes, and early discontinuation.6 Some trials do report the extent of these modifications, although uncommon in practice.7 There is some evidence from real-world studies to suggest that modifications such as early discontinuation, dose reductions and delays occur in 10% to 30% of patients.8–15 Notably, these outcomes have been examined individually, rather than in aggregate, and it remains unclear to what degree real-world patients receive chemotherapy as intended. This is important, as evidence suggests that successfully administering planned chemotherapy has been associated with improved survival,16 and treatment modifications, such as delays, are associated with excess resource utilization and increased healthcare costs.17

We describe the landscape of real-world breast cancer chemotherapy delivery in the community setting, and further report the extent of modifications overall, by drug combination, and for the most common specific administration schedules. We additionally examine how specific regimen characteristics relate to the likelihood of chemotherapy modifications.

Materials & Methods

The Optimal Breast Cancer Chemotherapy Dosing (OBCD) Study is a real-world cohort of 34,109 women diagnosed with primary stage I-IIIA breast cancer at Kaiser Permanente Northern California (KPNC) from 2006–2019 and Kaiser Permanente Washington (KPWA) from 2004–2015. Both are members of the Health Care Systems Research Network (HCSRN) and maintain a Virtual Data Warehouse, a common data model with standardized variable definitions, and data extracted from electronic health records (EHRs).18,19 Data were obtained from the VDW, EHRs and chart review.

Study participants were identified through the VDW based on the following criteria: primary breast cancer diagnosis with no prior history or same-day diagnosis of any cancer (except non-melanoma skin cancer), enrollment at KPWA or KPNC at time of diagnosis, had available medical records, and did not opt out of research studies. KPWA also included a subsample of women who were part of a previous study (COMBO, COmmonly used Medications and Breast Cancer Outcomes).20

Participants in this analysis were restricted to those with a first chemotherapy date after diagnosis, no missing chemotherapy data, and no known enrollment in RCTs (n=12,558) (Figure 1). Some analyses necessitated exclusion of those with an unknown intended regimen and non-guideline drug combination, as detailed below.

Figure 1:

Figure 1:

Open in a new tab

Study Inclusion Criteria

IRB approval was obtained from Memorial Sloan Kettering Cancer Center, Rutgers University, KPNC, and KPWA, with a waiver of consent at KPNC and KPWA.

Identifying the intended regimen

We identified the intended chemotherapy regimen through an iterative process, using algorithms and chart review, as detailed previously.21,22 Briefly, participants were categorized according to the drug combination received. Among those receiving drug combinations detailed in the National Comprehensive Cancer Network (NCCN) guidelines from 2004–2019, algorithms were used to classify intended administration schedule (cycle number and interval), with medical chart abstraction used where ambiguous. We additionally described non-guideline administration schedules and drug combinations. The intended regimen was identified through algorithms for 77% of participants, by chart review for 18%, and for 4.5%, the intended regimen could not be determined.

Regimens described

The ten most used drug combinations in this study are described in Figure 2, and their abbreviations are as follows:

Figure 2:

Figure 2:

Open in a new tab

Chemotherapy Regimen Descriptions

AC: cyclophosphamide, doxorubicin

AC+T(DOC): cyclophosphamide, doxorubicin, docetaxel

AC+T(PAC): cyclophosphamide, doxorubicin, paclitaxel

AC-T(PAC)-H: cyclophosphamide, doxorubicin, paclitaxel, trastuzumab

CMF: cyclophosphamide, methotrexate, 5-fluorouracil

HER-TC: cyclophosphamide, docetaxel, trastuzumab

TCH: docetaxel, carboplatin, trastuzumab

TCHP: docetaxel, carboplatin, trastuzumab, pertuzumab

TC: cyclophosphamide, docetaxel

TH: paclitaxel, trastuzumab

Modifications from intended chemotherapy

First cycle dose proportion (FCDP) <90%

The chemotherapy dose received was compared to that intended. Those who received <90%23,24 of the intended dose of their intended regimen, based on body surface area (BSA) using actual body weight, were classified as having a first cycle dose reduction.

Average relative dose intensity (ARDI) <90%

We calculated the relative dose intensity (RDI) for each drug, considering dosing and duration of received vs intended, and averaging this ratio across drugs to calculate the average relative dose intensity (ARDI), identifying those with an ARDI <90%,23,24 censored at the time of drug discontinuation. ARDIs were averaged for participants undergoing a regimen change. If participants received 3 or more regimens, chemotherapy was censored at the time of second regimen change.

Notably, oral drugs were not included in any dosing calculations.

Regimen change

Defined as a change in drug combination or administration schedule during treatment.

Fewer cycles than expected

We defined early discontinuation as unexpected cessation or administration of fewer than intended number of cycles, allowing for a 1-infusion buffer for trastuzumab/pertuzumab (which are most commonly intended for use over a “1-year period”).

Extended delay

Defined as any single delay of >14 days between cycles (excluding around surgery).

Modifications from intended chemotherapy

Modifications from intended chemotherapy comprised those who experienced any of the above.

Regimen features

We examined associations with modifications from intended chemotherapy, overall and for the constituent parts, by attributes of the intended regimen. Factors examined include NCCN-guideline concordance, dose density, any neoadjuvant/adjuvant only treatment, number of cycles (1–4; 5–10; 11–16; 17+), number of drugs (1–2; 3; 4+) and drug classes received as part of the intended regimen (anthracyclines; alkylating agents; antimetabolites; taxanes; monoclonal antibodies).

Statistical Analyses

We examined the study population, overall, and by receipt of chemotherapy modification, and have further described chemotherapy modifications for the 10 most common drug combinations (detailed in Figure 2). As a secondary analysis, we described the distribution of chemotherapy modifications for the 10 most common intended administration schedules within guideline drug combinations.

We examined associations of regimen attributes with modifications in chemotherapy using generalized linear models of the Poisson family with a log link-function and robust standard errors to calculate prevalence ratios (PR) and 95% Confidence Intervals (95% CI), adjusted for age at diagnosis (years; continuous), stage (I, II, IIIA), study site (KPNC, KPWA) and diagnosis year (2004–2010, 2011–2014, 2015–2019).

In this study, we calculated dosing based on actual body weight. However, given that the heterogeneity in approach on dosing for carboplatin, we conducted sensitivity analyses excluding carboplatin for the overall composite of chemotherapy modification, as well as both dosing variables (FCFP and ARDI).

Regimen identification algorithms were implemented in SAS v9.4, and Stata v17 was used for analyses.

Results

Overall, 12,558 women received chemotherapy, of whom 57.0% were non-Hispanic White, 8.1% non-Hispanic Black/African American, 17.9% Asian and 15.4% Hispanic (Table 1).

Table 1:

Population characteristics of those receiving chemotherapy overall and if chemotherapy was received as intended or modified from intended

Characteristic Total, those receiving chemotherapy Chemotherapy received as intended
Overall N = 12,558 (100%)1 Yes N = 7254 (64%)1 No N = 4016 (36%)1
Age at diagnosis (years)
 Median (IQR) 55 (47.0, 63) 54 (46.0, 62) 56 (48.0, 64)
 Mean (SD) 54.7 (11) 53.8 (11) 56.1 (11)
Race and ethnicity
 White 7,158 (57.0%) 4,050 (55.8%) 2,350 (58.5%)
 Black or African American 1,018 (8.1%) 544 (7.5%) 378 (9.4%)
 Asian 2,252 (17.9%) 1,381 (19.0%) 652 (16.2%)
 Hispanic 1,930 (15.4%) 1,163 (16.0%) 573 (14.3%)
 American Indian or Alaskan Native 40 (0.3%) 24 (0.3%) 13 (0.3%)
 Native Hawaiian or Other Pacific Islander 107 (0.9%) 66 (0.9%) 31 (0.8%)
 More than once race 24 (0.2%) 9 (0.1%) 11 (0.3%)
 Unknown/missing 29 (0.2%) 17 (0.2%) 8 (0.2%)
Year of diagnosis
 2004–2011 5,320 (42.4%) 2,746 (37.9%) 1,902 (47.4%)
 2012–2019 7,238 (57.6%) 4,508 (62.1%) 2,114 (52.6%)
BMI at diagnosis (kg/m 2 )
 Median (IQR) 27 (24.0, 33) 27 (23.8, 32) 29 (24.6, 34)
 Mean (SD) 28.8 (7) 28.2 (6) 30.1 (7)
 Missing 95 (0.8%) 5 (0.1%) 5 (0.1%)
Charlson comorbidity index
 0 9,479 (75.5%) 5,698 (78.5%) 2,838 (70.7%)
 1 1,996 (15.9%) 1,071 (14.8%) 733 (18.3%)
 2 601 (4.8%) 290 (4.0%) 233 (5.8%)
 3+ 482 (3.8%) 195 (2.7%) 212 (5.3%)
Grade
 Grade I 1,223 (9.7%) 773 (10.7%) 335 (8.3%)
 Grade II 5,182 (41.3%) 3,137 (43.2%) 1,560 (38.8%)
 Grade III/IV 5,795 (46.1%) 3,159 (43.5%) 1,985 (49.4%)
 Grade unknown, not stated, or not applicable 358 (2.9%) 185 (2.6%) 136 (3.4%)
Hormone-receptor status
 ER- and PR- 3,526 (28.1%) 1,891 (26.1%) 1,233 (30.7%)
 ER+ and/or PR+ 9,012 (71.8%) 5,351 (73.8%) 2,778 (69.2%)
 Borderline/not tested/unknown 20 (0.2%) 12 (0.2%) 5 (0.1%)
HER2 status
 HER2- 9,147 (72.8%) 5,931 (81.8%) 2,466 (61.4%)
 HER2+ 3,235 (25.8%) 1,240 (17.1%) 1,483 (36.9%)
 Borderline 97 (0.8%) 51 (0.7%) 38 (0.9%)
 Not tested/unknown 9 (0.6%) 32 (0.4%) 29 (0.7%)
Open in a new tab
1

n (%)

2

Final two columns use a denominator of 11,270, the subset for whom modifications to intended chemotherapy could be identified.

The top 10 drug combinations were received by 12,153 participants and 22 NCCN-guideline (2004–2019) regimens were used overall (Table 2).

Table 2 –

Chemotherapy modified from intended, by drug combination

Overall Neoadjuvant First Cycle Dose Proportion <90% ARDI <90% Regimen Change Fewer Infusions Than Expected a Extended Delay b Chemotherapy modified from intended, any reason
For all eligible participants: N = 11,270 (92%) N = 884 (7.8%) N = 726 (6.4%)1 N = 2,190 (19%)1 N = 470 (4.2%)1 N = 1,867 (17%)1 N = 915 (8.1%)1 N = 4,016 (36%)1
For top 10 drug combinations: N = 11,194 (92%) N = 864 (7.7%) N = 717 (6.4%)1 N = 2,166 (19%)1 N = 462 (4.1%)1 N = 1,838 (16%)1 N = 900 (8.0%)1 N = 3,965 (35%)1
Drug Combination
 AC 716 (6.4%) 15 (2.1%) 75 (10%) 185 (26%) 7 (1.0%) 67 (9.4%) 27 (3.8%) 237 (33%)
 AC+T(DOC) 286 (2.6%) 45 (16%) 43 (15%) 90 (31%) 13 (4.5%) 62 (22%) 25 (8.7%) 146 (51%)
 AC+T(PAC) 3,858 (34%) 426 (11%) 212 (5.5%) 704 (18%) 174 (4.5%) 618 (16%) 148 (3.8%) 1,312 (34%)
 AC+T(PAC)+H 567 (5.1%) 40 (7.1%) 31 (5.5%) 108 (19%) 56 (9.9%) 224 (40%) 181 (32%) 358 (63%)
 CMF 101 (0.9%) 0 (0%) 28 (28%) 42 (42%) 7 (6.9%) 17 (17%) 14 (14%) 54 (53%)
 HER-TC 79 (0.7%) 2 (2.5%) 5 (6.3%) 13 (16%) 2 (2.5%) 27 (34%) 15 (19%) 42 (53%)
 TCH 1,157 (10%) 44 (3.8%) 135 (12%) 412 (36%) 50 (4.3%) 297 (26%) 245 (21%) 664 (57%)
 TCHP 438 (3.9%) 233 (53%) 5 (1.1%) 125 (29%) 36 (8.2%) 124 (28%) 83 (19%) 235 (54%)
 TC 3,498 (31%) 55 (1.6%) 161 (4.6%) 392 (11%) 113 (3.2%) 271 (7.7%) 73 (2.1%) 716 (20%)
 TH 494 (4.4%) 4 (0.8%) 22 (4.5%) 95 (19%) 4 (0.8%) 131 (27%) 89 (18%) 201 (41%)
Total 11,194 (100%) 864 (7.7%) 717 (6.4%) 2,166 (19%) 462 (4.1%) 1,838 (16%) 900 (8.0%) 3,965 (35%)
Open in a new tab
1

Percent is out of total able to calculate outcome variables

a

Fewer cycles than expected: unexpected cessation of a single agent or administration of fewer than intended number of cycles; for patients receiving TC and for whom chart abstraction revealed an intent to administer 4–6 cycles, we have grouped these patients with patients intended to receive TCx6, and have not considered them to receive fewer cycles than expected if 4 or 5 cycles ultimately are received

b

Extended delay: Patients were described as experiencing an extended delay if they had any single delay of >14 days between cycles (excluding delays around surgery)

If the intended regimen could be determined, we were then able to assess if chemotherapy modifications occurred; this applied to 11,270 participants or 92% of the original cohort (defined as “Table 2 subcohort”), among whom 36% (n=4,016) did not receive chemotherapy as intended. All numbers hereafter refer to the Table 2 subcohort.

Of the top 10 drug combinations, those with the highest proportions not receiving chemotherapy as intended were AC+T(PAC)+H (63%), TCH (57%) and TCHP (54%). FCDP<90% was identified in 6.4% and was most common in people receiving CMF (28%). ARDI<90% was present in 19%, and most prevalent in people receiving CMF (42%), TCH (36%) and AC+T(DOC) (31%). Regimen change was observed in 4.2%. The most common regimen with fewer infusions than expected (40%) and extended delays (32%) was AC+T(PAC)+H. Neoadjuvant chemotherapy was received by 7.8% of participants and was most often observed in TCHP (53%).

The most common specific administration schedules are delineated in Figure 1. Of these most common administration schedules, both TCHP Regimen A and TCH Regimen A had >50% who did not receive chemotherapy as intended, with three ACT administration schedules also seeing substantive chemotherapy modification (Supplemental Table 1). An ARDI<90% was more common in ACT-A and ACT-B (both of which were not dose-dense), than ACT-D (dose-dense), although 18% of ACT-D received fewer cycles than expected (compared to <10% of ACT-A and ACT-B)

We examined associations between regimen characteristics and chemotherapy modifications (Table 3). Those intended to receive guideline-concordant chemotherapy were less likely to have modifications (PR:0.74; 95% CI:0.69–0.79); specifically with a lower likelihood of extended delay (PR:0.46; 95% CI:0.40–0.54) and fewer infusions than expected (PR:0.71; 95%CI:0.63–0.80). Those receiving dose-dense regimens were less likely than those receiving non-dose dense regimens to experience chemotherapy modifications (PR:0.84; 95% CI:0.78–0.91). Those receiving any neoadjuvant vs adjuvant chemotherapy only chemotherapy were more likely to experience modifications (PR:1.29; 95%CI:1.19–1.41), with an increased likelihood of regimen change (PR:1.94; 95%CI:1.50–2.52) and receipt of fewer infusions than expected (PR:1.45; 95%CI:1.27–1.66).

Table 3 –

Association between chemotherapy features and chemotherapy modified from intended, PR and 95% CI

First Cycle Dose Reduction <90% Average Relative Dose Intensity <90% Extended Delay b Regimen Change Fewer Cycles Than Expected a Chemotherapy Modified from Intended, Any reason
Total n (%) n (%) PR (95% CI) n (%) PR (95% CI) n (%) PR (95% CI) n (%) PR (95% CI) n (%) PR (95% CI) n (%) PR (95% CI)
Characteristic of Regimen
NCCN Guideline
 Non guideline 1,116 (9.9) 94 (13.0) REF 247 (11.3) REF 179 (19.6) REF 89 (18.9) REF 257 (13.8) REF 525 (13.1) REF
 Guideline 10,154 (90.1) 632 (87.0) 0.75 (0.61, 0.91) 1,943 (88.7) 0.88 (0.78, 0.98) 736 (80.4) 0.46 (0.40, 0.54) 381 (81.1) 0.46 (0.37, 0.58) 1,610 (86.2) 0.71 (0.63, 0.80) 3,491 (86.9) 0.74 (0.69, 0.79)
Dose Dense Regimen among those with dose dense option
 Not dose dense 2166 (39.6) 232 (63.2) REF 541 (49.1) REF 186 (47.3) REF 70 (27.7) REF 467 (47.0) REF 948 (45.5) REF
 Dose dense 3307 (60.4) 135 (36.8) 0.51 (0.40, 0.63) 561 (50.9) 0.79 (0.70, 0.88) 207 (52.7) 0.72 (0.58, 0.90) 183 (72.3) 2.25 (1.67, 3.04) 527 (53.0) 0.71 (0.63, 0.81) 1,138 (54.5) 0.84 (0.78, 0.91)
Chemotherapy timing
 Adjuvant 10,386 (92.2) 683 (94.1) REF 2,015 (92.0) REF 852 (93.1) REF 407 (86.6) REF 1,672 (89.6) REF 3,642 (90.7) REF
 Neoadjuvant 884 (7.8) 43 (5.9) 1.06 (0.79, 1.44) 175 (8.0) 1.12 (0.98, 1.29) 63 (6.9) 0.89 (0.69, 1.14) 63 (13.4) 1.94 (1.50, 2.52) 195 (10.4) 1.45 (1.27, 1.66) 374 (9.3) 1.29 (1.19, 1.41)
Drug Class
Anthracycline
 No 5,767 (51.2) 356 (49.0) REF 1,079 (49.3) REF 519 (56.7) REF 212 (45.1) REF 867 (46.4) REF 1,912 (47.6) REF
 Yes 5,503 (48.8) 370 (51.0) 0.92 (0.78, 1.08) 1,111 (50.7) 1.03 (0.94, 1.12) 396 (43.3) 0.70 (0.61, 0.80) 258 (54.9) 1.01 (0.83, 1.24) 1,000 (53.6) 1.24 (1.13, 1.37) 2,104 (52.4) 1.11 (1.05, 1.18)
Taxane
 No 2,433 (21.6) 245 (33.8) REF 771 (35.2) REF 371 (40.6) REF 102 (21.7) REF 508 (27.2) REF 1,202 (29.9) REF
 Yes 8,837 (78.4) 481 (66.3) 0.59 (0.51, 0.69) 1,419 (64.8) 0.51 (0.47, 0.55) 544 (59.4) 0.38 (0.34, 0.44) 368 (78.3) 0.97 (0.78, 1.21) 1,359 (72.8) 0.72 (0.65, 0.79) 2,814 (70.1) 0.64 (0.60, 0.67)
Monoclonal antibody
 No 8,487 (75.3) 523 (72.0) REF 1,424 (65.0) REF 290 (31.7) REF 318 (67.7) REF 1,041 (55.8) REF 2,482 (61.8) REF
 Yes 2,783 (24.7) 203 (28.0) 1.39 (1.19, 1.62) 766 (35.0) 1.75 (1.62, 1.89) 625 (68.3) 7.12 (6.23, 8.14) 152 (32.3) 1.64 (1.36, 1.98) 826 (44.2) 2.57 (2.37, 2.78) 1,534 (38.2) 2.01 (1.92, 2.11)
Alkylating Agent
 No 494 (4.4) 22 (3.0) REF 95 (4.3) REF 89 (9.7) REF 4 (0.9) REF 131 (7.0) REF 201 (5.0) REF
 Yes 10,776 (95.6) 704 (97.0) 1.03 (0.67, 1.58) 2,095 (95.7) 0.95 (0.78, 1.15) 826 (90.3) 0.35 (0.28, 0.44) 466 (99.2) 3.65 (1.35, 9.84) 1,736 (93.0) 0.57 (0.48, 0.67) 3,815 (95.0) 0.79 (0.70, 0.89)
Antimetabolite
 No 11,136 (98.8) 694 (95.6) REF 2,137 (97.6) REF 897 (98.0) REF 458 (97.5) REF 1,841 (98.6) REF 3,941 (98.1) REF
 Yes 134 (1.2) 32 (4.4) 2.21 (1.64, 2.98) 53 (2.4) 1.64 (1.33, 2.03) 18 (2.0) 1.57 (1.01, 2.43) 12 (2.6) 1.95 (1.12, 3.37) 26 (1.4) 0.99 (0.70, 1.40) 75 (1.9) 1.33 (1.14, 1.55)
Number of cycles
 1–4 3,906 (34.7) 217 (29.9) REF 528 (24.1) REF 86 (9.4) REF 104 (22.2) REF 300 (16.1) REF 859 (21.4) REF
 5–10 2329 (20.7) 206 (28.4) 1.48 (1.21, 1.80) 465 (21.2) 1.41 (1.25, 1.59) 87 (9.5) 1.44 (1.07, 1.96) 163 (34.7) 2.21 (1.70, 2.86) 269 (14.4) 1.41 (1.20, 1.66) 786 (19.6) 1.43 (1.32, 1.56)
 11–16 2258 (20.0) 103 (14.2) 1.18 (0.93, 1.50) 432 (19.7) 1.61 (1.43, 1.82) 116 (12.7) 2.38 (1.80, 3.16) 49 (10.4) 0.81 (0.57, 1.15) 475 (25.4) 3.01 (2.61, 3.46) 840 (20.9) 1.85 (1.71, 2.01)
 17+ 2777 (24.6) 200 (27.6) 1.59 (1.32, 1.92) 765 (34.9) 2.20 (2.00, 2.43) 626 (68.4) 10.64 (8.53, 13.27) 154 (32.8) 2.14 (1.67, 2.73) 823 (44.1) 4.12 (3.64, 4.65) 1,531 (38.1) 2.67 (2.50, 2.85)
p-trend < 0.001 p-tend < 0.001 p-trend < 0.001 p-trend < 0.001 p-trend < 0.001 p-trend < 0.001
Number of drugs
 1–2 4,708 (41.8) 258 (35.5) REF 672 (30.7) REF 189 (20.7) REF 124 (26.4) REF 469 (25.1) REF 1,154 (28.7) REF
 3 5,502 (48.8) 425 (58.5) 1.53 (1.30, 1.79) 1,268 (57.9) 1.75 (1.60, 1.92) 449 (49.1) 2.31 (1.94, 2.74) 248 (52.8) 1.63 (1.30, 2.04) 1,024 (54.9) 2.08 (1.87, 2.31) 2,230 (55.5) 1.78 (1.67, 1.89)
 4+ 1,060 (9.4) 43 (5.9) 0.92 (0.67, 1.26) 250 (11.4) 1.89 (1.65, 2.16) 277 (30.3) 7.71 (6.44, 9.23) 98 (20.9) 3.49 (2.66, 4.56) 374 (20.0) 4.21 (3.72, 4.76) 632 (15.7) 2.75 (2.56, 2.96)
p-trend = 0.004 p-trend < 0.001 p-trend < 0.001 p-trend < 0.001 p-trend < 0.001 p-trend < 0.001
Open in a new tab

*adjusted for age at diagnosis, stage, and years of diagnosis, and health system site

a

Fewer cycles than expected: unexpected cessation of a single agent or administration of fewer than intended number of cycles; for patients receiving TC and for whom chart abstraction revealed an intent to administer 4–6 cycles, we have grouped these patients with patients intended to receive TCx6, and have not considered them to receive fewer cycles than expected if 4 or 5 cycles ultimately are received

b

Extended delay: Patients were described as experiencing an extended delay if they had any single delay of >14 days between cycles (excluding delays around surgery)

Drug class was also associated with chemotherapy modifications (Table 3). Persons receiving anthracyclines were more likely to receive fewer infusions than expected (PR:1.24; 95%CI:1.13–1.37), although were less likely to experience delays (PR:0.70; 95% CI:0.61–0.80). Both taxanes (PR:0.64; 95%CI: 0.60–0.67) and alkylating agents (PR:0.79; 95%CI:0.70–0.89) were associated with a lower likelihood of chemotherapy modifications. Those receiving taxanes had lower likelihood of dose reductions, extended delay, and receiving fewer infusions than expected. Those receiving alkylating agents were less likely to have delays (PR:0.35; 95%CI:0.28–0.44) and fewer cycles than expected (PR:0.57; 95%CI:0.48–0.67) but more likely to experience regimen change (PR:3.65; 95%CI:1.35–9.84). The use of monoclonal antibodies was positively associated with all modifications, including markedly higher likelihood of extended delay (PR:7.12; 95%CI:6.23–8.14) and receiving fewer infusions than expected (PR:2.57; 95%CI:2.37–2.78).

Patients receiving regimens with more cycles were more likely to experience chemotherapy modification (PR:17+ cycles vs ≤4 cycles:2.67; 95% CI:2.50–2.85)(p-trend<0.001). Increasing number of drugs was also associated (PR:4+ drugs vs 1–2:2.75; 95%CI:2.56–2.96)(p-trend<0.001).

In sensitivity analyses excluding carboplatin, the percentage receiving a modification decreased slightly, and results were largely unchanged. While results for the overall composite of chemotherapy modification remained unchanged, as did results for ARDI, a few significant associations for first cycle dose reductions attenuated and became non-significant, namely those variables expected to be impacted directly in the HER2+ group receiving the carboplatin-based regimens, including number of cycles, number of drugs, and regimens containing monoclonal antibodies – in line with the recommended use of carboplatin in HER2+ patients (Supplemental Table 2).

Discussion

In this comprehensive assessment of real-world use of chemotherapy in breast cancer, approximately 36% of patients did not receive treatments as intended. Modifications from intended chemotherapy varied widely by regimen administered, at both drug combination and administration-schedule level. Receipt of neoadjuvant chemotherapy, anthracyclines, monoclonal antibodies and antimetabolites, and an increased number of cycles and drugs were associated with increased likelihood of chemotherapy modifications. This is important, as modifications impact the ability to infer real-world effectiveness for regimens which are efficacious in the trial setting due to differences in real-world delivery.17,25

To our knowledge, data on modifications in chemotherapy by drug regimen and administration schedule have not previously been reported within a single comprehensive study. The common drug combinations with the highest percentage of people receiving chemotherapy modifications were AC+T(PAC)-H, TCH, and TCHP, which notably all include a minimum of three drugs, one of which is trastuzumab. Trastuzumab is given over the course of a 1-year period (longer than other drugs recommended by the NCCN), providing more opportunity for modifications in care. Obtaining directly comparable data is challenging, and different trial protocols recommend varying responses to toxicity; some include dose reductions with no re-escalation and some include BSA cut-offs for obese women (common among historical trials published prior to the 2012 American Society of Clinical Oncology guidelines which recommended full-BSA dosing).26 Even so, an analysis of two AC+T(PAC)-H trials found that 31.4% of participants did not complete trastuzumab, which was lower than observed in this study (we found that 40% of participants receiving AC+T(PAC)-H had fewer cycles than expected, largely driven by trastuzumab);27,28 across trials, 14.2% of non-completion was due to decline in left ventricular ejection fraction.15,27–29 In both trials, if trastuzumab was held for toxicity, the protocol proposed completing treatment at one year, regardless of how many doses had been omitted. This practice may explain the high proportion with extended delays and who received fewer cycles than intended.

Of the top 10 drug combinations, TC had the lowest proportion of patients with modifications from the intended regimen. TC has been well-tolerated in trials,30 and is generally preferred in older, HER2- patients.31 TC is also a short regimen, involving only 2 drugs over 4–6 cycles in practice. This may provide fewer opportunities for modifications in care. During medical chart abstraction, a small minority of TC patients were intended to receive “4–6” cycles. We did not classify these as early discontinuations if 4 or more cycles were received, which may also have contributed to this finding.

Patients receiving dose-dense chemotherapy were less likely to experience modifications from intended chemotherapy compared to non-dose-dense regimens, as observed in previous studies.32 Although dose-density is associated with myelosuppression, the use of granulocyte-stimulating colony factors (G-CSF) may improve the toxicity profile of dose-dense regimens and thus improve treatment completion.33 Patients receiving dose-dense regimens may be younger with fewer comorbidities, which may lower their likelihood of chemotherapy modifications.

Patients receiving guideline-concordant administration schedules were less likely to receive modifications as compared to patients receiving guideline-discordant administration schedules- possibly reflecting differing patient factors between those receiving guideline/non-guideline regimens, including comorbidity, frailty, or socioeconomic factors that affect treatment scheduling, such as employment and caregiving responsibilities.34–37

Those receiving any neoadjuvant chemotherapy were more likely to have modifications to intended chemotherapy compared to adjuvant chemotherapy only. To our knowledge, no other studies have observed this association. This may be due to delayed or omitted doses or changes in regimen to complete chemotherapy around surgical scheduling. Notably, delays to surgery in patients undergoing neoadjuvant chemotherapy are adversely associated with survival,38 and identifying regimens most likely to experience delays may be helpful in planning neoadjuvant chemotherapy.

An increased likelihood of delays was found in patients receiving monoclonal antibodies, anthracyclines and antimetabolites. Anti-HER-2 therapy is associated with cardiotoxicity, which may lead to treatment delays, and the long duration of monoclonal antibody administration may provide more opportunity for schedule modification. Anthracyclines are associated with substantial toxicity, potentially impacting modifications.39 To this end, Liutkauskiene et al13 found that neutropenia and non-medical reasons (including patient request and logistical challenges) were the main drivers of anthracycline dose reductions. Regardless, prior studies have shown dose reductions and delays in anthracyclines to be associated with poorer survival.13,40 The most common antimetabolite used in this cohort was 5-fluorouracil (5FU), which is associated with a dose-limiting toxicity profile, including myelosuppression, mucositis and severe gastrointestinal effects.41,42 Dose, route and schedule of 5-FU affect pharmacokinetics and treatment effects,43,44 suggesting modifications in 5FU may materially impact outcomes. Those receiving taxanes and alkylating agents were less likely to have modifications; these drugs cause less toxicity than anthracyclines and monoclonal antibodies.45,46

A higher likelihood of modifications was observed with increasing number of cycles, which is intuitive, given increased opportunity for modifications. Modifications were more common in administration schedules AC-PAC(T)-A and AC-PAC(T)-B, which are both non-dose dense anthracycline-containing regimens and the reasons outlined above may drive these associations.

Although we examined detailed chemotherapy data in a large population, it is possible some intended regimens were misclassified. We mitigated this with chart abstraction in cases with ambiguity. As with any observational study, confounding may have been possible. We examined associations by guideline-concordance using NCCN guidelines, which may not always be referenced by prescribing oncologists, although are aligned with other global guidelines. Associations with alkylating agents may have been affected by carboplatin, for which there may be more heterogeneity in approach on dosing; in this study, we calculated dosing based on actual body weight. In sensitivity analyses excluding carboplatin from dosing calculations (ARDI and FCDP). we found a small decrease in the percent receiving ARDI<90% and FCDP<90%; while results for the main outcome of chemotherapy modification and ARDI were comparable, we provide results, given the change in some variables associated with the FCDP component. Ranges are recommended for some drugs in NCCN guidelines; while we calculated conservatively based on the lower end of the range, associations may have varied depending on the reference dose used. Furthermore, this study does not distinguish those with first cycle dose reductions where an escalation was intended at the outset. Therefore, first cycle dose reductions may conceivably include some patients who subsequently experienced dose escalation. Even so, dose escalation remains relatively infrequent in the cohort. For chemotherapy characteristics in relation to modification, it was not possible to present mutually-adjusted models given the strong correlations between chemotherapy attributes, such as number of cycles and use of monoclonal antibodies. However, associations for number of cycles restricted to non-trastuzumab-containing regimens were comparable to initial findings (results not shown). Additionally, confounding by indication or comorbidity may drive some of these associations, e.g. where CMF is used where anthracyclines/taxanes are contraindicated. Regardless, we have elected the approach taken to reflect what occurs in real-world practice, providing a springboard for further research. The patient factors driving some of these associations in this population has been commented on elsewhere.47 Even so, we are unable to comment here on the specific precipitating issue behind these modifications, whether that be an issue with tolerability, toxicity, scheduling, or otherwise. Our findings of significant heterogeneity in practice have been observed for different treatment-related outcomes in this population, including treatment delays, non-guideline regimen use and dose reductions.23,48,49 Our study is conducted in an integrated healthcare setting, and one must consider generalizability across healthcare systems. Even so, research has shown this setting demographically similar to its underlying population and this setting makes it uniquely possible to study this question over and above insurance coverage.50 Lastly, given that our study ended in 2019, future studies may seek to expand this work to include newer treatment options, including new drugs and immunotherapy-based regimens.

In the first study of its kind, we leveraged a rich, large dataset on real-world chemotherapy use to describe modifications from intended chemotherapy by regimen. This identifies regimens with limits for inferring real-world effectiveness, despite efficacy in trials, and regimens which would benefit from further research to best serve the real-world patients receiving them.

In summary, we observe extensive heterogeneity in chemotherapy delivery and identify specific regimen characteristics that may prompt further research on their real-world effectiveness. Chemotherapy modifications varied widely dependent on the regimen administered, in both type of modification and likelihood of modifications occurring. Factors such as increased number of cycles and drugs, receipt of neoadjuvant chemotherapy were associated increased likelihood of chemotherapy modifications, as were regimens containing certain drug classes. These novel findings are valuable, not only to inform future research, but to facilitate planning around chemotherapy delivery in the community setting and improve patient outcomes.

Supplementary Material

Supplemental Table 1
Supplemental Table 2

Funding:

This work was supported by grants R37CA222793, U24CA171524, U01CA195565 and P30CA008748 from the National Cancer Institute of the National Institutes of Health, as well as the Geoffrey Beene Cancer Research Center at Memorial Sloan Kettering Cancer Center. Erin Bowles’s time was supported by the National Cancer Institute (R50CA211115).

Dr. Bandera served as member of Pfizer’s Advisory Board to enhance minority participation in clinical trials (7/2021-8/2023). Dr. Liu served as member of Pfizer’s think tank on real-world evidence sponsored by Pfizer on 11/28/23. Dr. Liu’s research funding (inst) unrelated to this current work: Genentech, AstraZeneca, Bristol Myers Squibb, Exact Sciences, Biotheranostics, and Beigene.

Footnotes

Conflict of interest statement: Other authors have no COI to declare.

Data Availability Statement:

Data contain potentially identifiable information such as dates of events that cannot be shared openly without appropriate human subjects approval and data use agreements, but the data that support the findings of this study are available from the corresponding author upon request.

References

  • 1.Blonde L, Khunti K, Harris SB, Meizinger C, Skolnik NS. Interpretation and Impact of Real-World Clinical Data for the Practicing Clinician. Adv Ther. 2018;35(11):1763–1774. doi: 10.1007/s12325-018-0805-y [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Cohen AT, Goto S, Schreiber K, Torp-Pedersen C. Why do we need observational studies of everyday patients in the real-life setting? Eur Heart J Suppl. 2015;17(suppl_D):D2–D8. doi: 10.1093/eurheartj/suv035 [DOI] [Google Scholar]
  • 3.Barriers to Clinical Trial Enrollment in Racial and Ethnic Minority Patients with Cancer - Lauren M. Hamel, Louis A. Penner, Terrance L. Albrecht, Elisabeth Heath, Clement K. Gwede, Susan Eggly, 2016. Accessed January 19, 2024. https://journals.sagepub.com/doi/abs/10.1177/107327481602300404 [DOI] [PMC free article] [PubMed]
  • 4.Garg P, Rana F, Gupta R, Buzaianu EM, Guthrie TH. Predictors of Toxicity and Toxicity Profile of Adjuvant Chemotherapy in Elderly Breast Cancer Patients. The Breast Journal. 2009;15(4):404–408. doi: 10.1111/j.1524-4741.2009.00745.x [DOI] [PubMed] [Google Scholar]
  • 5.Freedman RA, Vaz-Luis I, Barry WT, et al. Patterns of chemotherapy, toxicity, and short-term outcomes for older women receiving adjuvant trastuzumab-based therapy. Breast Cancer Res Treat. 2014;145(2):491–501. doi: 10.1007/s10549-014-2968-9 [DOI] [PubMed] [Google Scholar]
  • 6.Largent EA, Lynch HF. Addressing Financial Barriers to Enrollment in Clinical Trials. JAMA Oncol. 2018;4(7):913–914. doi: 10.1001/jamaoncol.2018.0492 [DOI] [PubMed] [Google Scholar]
  • 7.Eichler M, Singer S, Janni W, et al. Pretreatment quality of life, performance status and their relation to treatment discontinuation and treatment changes in high-risk breast cancer patients receiving chemotherapy: results from the prospective randomized ADEBAR trial. Breast Cancer. 2017;24(2):319–325. doi: 10.1007/s12282-016-0706-3 [DOI] [PubMed] [Google Scholar]
  • 8.Aleixo GFP, Shachar SS, Deal AM, et al. The association of body composition parameters and adverse events in women receiving chemotherapy for early breast cancer. Breast Cancer Res Treat. 2020;182(3):631–642. doi: 10.1007/s10549-020-05731-1 [DOI] [PubMed] [Google Scholar]
  • 9.Neugut AI, Hillyer GC, Kushi LH, et al. A prospective cohort study of early discontinuation of adjuvant chemotherapy in women with breast cancer: the breast cancer quality of care study (BQUAL). Breast Cancer Res Treat. 2016;158(1):127–138. doi: 10.1007/s10549-016-3855-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Neugut AI, Hillyer GC, Kushi LH, et al. Non-initiation and early discontinuation of adjuvant trastuzumab in women with localized HER2-positive breast cancer. Breast Cancer. 2014;21(6):780–785. doi: 10.1007/s12282-014-0543-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Adjogatse D, Thanopoulou E, Okines A, et al. Febrile Neutropaenia and Chemotherapy Discontinuation in Women Aged 70 Years or Older Receiving Adjuvant Chemotherapy for Early Breast Cancer. Clinical Oncology. 2014;26(11):692–696. doi: 10.1016/j.clon.2014.05.002 [DOI] [PubMed] [Google Scholar]
  • 12.Zanuso V, Fregoni V, Gervaso L. Side effects of adjuvant chemotherapy and their impact on outcome in elderly breast cancer patients: a cohort study. Future Science OA. 2020;6(9):FSO617. doi: 10.2144/fsoa-2020-0076 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Liutkauskiene S, Grizas S, Jureniene K, Suipyte J, Statnickaite A, Juozaityte E. Retrospective analysis of the impact of anthracycline dose reduction and chemotherapy delays on the outcomes of early breast cancer molecular subtypes. BMC Cancer. 2018;18(1):1–9. doi: 10.1186/s12885-018-4365-y [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Shayne M, Crawford J, Dale DC, Culakova E, Lyman GH, for the ANC Study Group. Predictors of reduced dose intensity in patients with early-stage breast cancer receiving adjuvant chemotherapy. Breast Cancer Res Treat. 2006;100(3):255–262. doi: 10.1007/s10549-006-9254-4 [DOI] [PubMed] [Google Scholar]
  • 15.Griggs JJ, Sorbero MES, Stark AT, Heininger SE, Dick AW. Racial Disparity in the Dose and Dose Intensity of Breast Cancer Adjuvant Chemotherapy. Breast Cancer Res Treat. 2003;81(1):21–31. doi: 10.1023/A:1025481505537 [DOI] [PubMed] [Google Scholar]
  • 16.O’Connor TL, Edge SB, Kossoff EB, et al. Factors affecting the delivery of adjuvant/neoadjuvant chemotherapy in older women with breast cancer. Journal of Geriatric Oncology. 2012;3(4):320–328. doi: 10.1016/j.jgo.2012.06.006 [DOI] [Google Scholar]
  • 17.Elting LS, Rubenstein EB, Martin CG, et al. Incidence, Cost, and Outcomes of Bleeding and Chemotherapy Dose Modification Among Solid Tumor Patients With Chemotherapy-Induced Thrombocytopenia. Journal of Clinical Oncology. Published online September 21, 2016. doi: 10.1200/JCO.2001.19.4.1137 [DOI] [PubMed] [Google Scholar]
  • 18.Ross TR, Ng D, Brown JS, et al. The HMO Research Network Virtual Data Warehouse: A Public Data Model to Support Collaboration. EGEMS (Wash DC). 2014;2(1):1049. doi: 10.13063/2327-9214.1049 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Chubak J, Ziebell R, Greenlee RT, et al. The Cancer Research Network: a platform for epidemiologic and health services research on cancer prevention, care, and outcomes in large, stable populations. Cancer Causes Control. 2016;27(11):1315–1323. doi: 10.1007/s10552-016-0808-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Boudreau DM, Yu O, Chubak J, et al. Comparative safety of cardiovascular medication use and breast cancer outcomes among women with early stage breast cancer. Breast Cancer Res Treat. 2014;144(2):405–416. doi: 10.1007/s10549-014-2870-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Bhimani J, O’Connell K, Ergas I, et al. A methodology for using real-world data from electronic health records to assess chemotherapy administration in women with breast cancer. JCO CCI. Published online in press. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Bhimani J, O’Connell K, Ergas IJ, et al. Trends in chemotherapy use for early-stage breast cancer from 2006 to 2019. Breast Cancer Res. 2024;26(1):101. doi: 10.1186/s13058-024-01822-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Kantor ED, O’Connell K, Ergas IJ, et al. Assessment of breast cancer chemotherapy dose reduction in an integrated healthcare delivery system. Breast Cancer Res Treat. Published online November 4, 2023:1–10. doi: 10.1007/s10549-023-07126-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Griggs JJ, Sorbero MES, Lyman GH. Undertreatment of Obese Women Receiving Breast Cancer Chemotherapy. Arch Intern Med. 2005;165(11):1267–1273. doi: 10.1001/archinte.165.11.1267 [DOI] [PubMed] [Google Scholar]
  • 25.Weycker D, Edelsberg J, Kartashov A, Barron R, Lyman G. Risk and Healthcare Costs of Chemotherapy-Induced Neutropenic Complications in Women with Metastatic Breast Cancer. Chemotherapy. 2012;58(1):8–18. doi: 10.1159/000335604 [DOI] [PubMed] [Google Scholar]
  • 26.Griggs JJ, Mangu PB, Anderson H, Balaban EP, Dignam JJ, Hryniuk WM, Morrison VA, Pini TM, Runowicz CD, Rosner GL, Shayne M, Sparreboom A, Sucheston LE, Lyman GH; American Society of Clinical Oncology. Appropriate chemotherapy dosing for obese adult patients with cancer: American Society of Clinical Oncology clinical practice guideline. J Clin Oncol. 2012. May 1;30(13):1553–61. doi: 10.1200/JCO.2011.39.9436. Epub 2012 Apr 2. [DOI] [PubMed] [Google Scholar]
  • 27.Slamon D, Eiermann W, Robert N, et al. Adjuvant Trastuzumab in HER2-Positive Breast Cancer. New England Journal of Medicine. 2011;365(14):1273–1283. doi: 10.1056/NEJMoa0910383 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Vargas CE, Thorpe CS, Dueck AC, et al. Impact of adjuvant trastuzumab on locoregional failure rates in a randomized clinical trial: North Central Cancer Treatment Group N9831 (alliance) study. Cancer. 2020;126(24):5239–5246. doi: 10.1002/cncr.33154 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Romond EH, Perez EA, Bryant J, et al. Trastuzumab plus Adjuvant Chemotherapy for Operable HER2-Positive Breast Cancer. New England Journal of Medicine. 2005;353(16):1673–1684. doi: 10.1056/NEJMoa052122 [DOI] [PubMed] [Google Scholar]
  • 30.Jones S, Holmes FA, O’Shaughnessy J, et al. Docetaxel With Cyclophosphamide Is Associated With an Overall Survival Benefit Compared With Doxorubicin and Cyclophosphamide: 7-Year Follow-Up of US Oncology Research Trial 9735. Journal of Clinical Oncology. Published online February 9, 2009. doi: 10.1200/JCO.2008.18.4028 [DOI] [PubMed] [Google Scholar]
  • 31.Freedman RA, Li T, Sedrak MS, et al. ‘ADVANCE’ (a pilot trial) ADjuVANt chemotherapy in the elderly: Developing and evaluating lower-toxicity chemotherapy options for older patients with breast cancer. Journal of Geriatric Oncology. 2023;14(1):101377. doi: 10.1016/j.jgo.2022.09.006 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Venturini M, Del Mastro L, Aitini E, et al. Dose-Dense Adjuvant Chemotherapy in Early Breast Cancer Patients: Results From a Randomized Trial. J Natl Cancer Inst. 2005;97(23):1724–1733. doi: 10.1093/jnci/dji398 [DOI] [PubMed] [Google Scholar]
  • 33.Lyman GH. Impact of Chemotherapy Dose Intensity on Cancer Patient Outcomes. Journal of the National Comprehensive Cancer Network. 2009;7(1):99–108. doi: 10.6004/jnccn.2009.0009 [DOI] [PubMed] [Google Scholar]
  • 34.Griggs JJ, Culakova E, Sorbero ME, et al. Social and racial disparities in the use of non-standard breast cancer adjuvant chemotherapy regimens. Journal of Clinical Oncology. Published online June 20, 2006. doi: 10.1200/jco.2006.24.18_suppl.6036 [DOI] [Google Scholar]
  • 35.Griggs JJ, Culakova E, Sorbero MES, et al. Social and racial differences in selection of breast cancer adjuvant chemotherapy regimens. J Clin Oncol. 2007;25(18):2522–2527. doi: 10.1200/JCO.2006.10.2749 [DOI] [PubMed] [Google Scholar]
  • 36.Arts DL, Voncken AG, Medlock S, Abu-Hanna A, van Weert HCPM. Reasons for intentional guideline non-adherence: A systematic review. International Journal of Medical Informatics. 2016;89:55–62. doi: 10.1016/j.ijmedinf.2016.02.009 [DOI] [PubMed] [Google Scholar]
  • 37.Heins MJ, de Jong JD, Spronk I, Ho VKY, Brink M, Korevaar JC. Adherence to cancer treatment guidelines: influence of general and cancer-specific guideline characteristics. European Journal of Public Health. 2017;27(4):616–620. doi: 10.1093/eurpub/ckw234 [DOI] [PubMed] [Google Scholar]
  • 38.de Melo Gagliato D, Lei X, Giordano SH, et al. Impact of Delayed Neoadjuvant Systemic Chemotherapy on Overall Survival Among Patients with Breast Cancer. The Oncologist. 2020;25(9):749–757. doi: 10.1634/theoncologist.2019-0744 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Greene J, Hennessy B. The role of anthracyclines in the treatment of early breast cancer. Journal of Oncology Pharmacy Practice. Published online April 26, 2014. doi: 10.1177/1078155214531513 [DOI] [PubMed] [Google Scholar]
  • 40.Chirivella I, Bermejo B, Insa A, et al. Optimal delivery of anthracycline-based chemotherapy in the adjuvant setting improves outcome of breast cancer patients. Breast Cancer Res Treat. 2009;114(3):479–484. doi: 10.1007/s10549-008-0018-1 [DOI] [PubMed] [Google Scholar]
  • 41.Garg MB, Lincz LF, Adler K, Scorgie FE, Ackland SP, Sakoff JA. Predicting 5-fluorouracil toxicity in colorectal cancer patients from peripheral blood cell telomere length: a multivariate analysis. Br J Cancer. 2012;107(9):1525–1533. doi: 10.1038/bjc.2012.421 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Boussios S, Pentheroudakis G, Katsanos K, Pavlidis N. Systemic treatment-induced gastrointestinal toxicity: incidence, clinical presentation and management. Annals of Gastroenterology. 2012;25(2):106. [PMC free article] [PubMed] [Google Scholar]
  • 43.Dose and time dependencies of 5-fluorouracil pharmacokinetics - Terret - 2000 - Clinical Pharmacology & Therapeutics - Wiley Online Library. Accessed November 30, 2023. https://ascpt.onlinelibrary.wiley.com/doi/full/10.1067/mcp.2000.109352 [DOI] [PubMed]
  • 44.Kosmas C, Kallistratos MS, Kopterides P, et al. Cardiotoxicity of fluoropyrimidines in different schedules of administration: a prospective study. J Cancer Res Clin Oncol. 2008;134(1):75–82. doi: 10.1007/s00432-007-0250-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45.Delgado-Ramos GM, Nasir SS, Wang J, Schwartzberg LS. Real-world evaluation of effectiveness and tolerance of chemotherapy for early-stage breast cancer in older women. Breast Cancer Res Treat. 2020;182(2):247–258. doi: 10.1007/s10549-020-05684-5 [DOI] [PubMed] [Google Scholar]
  • 46.Du XL, Osborne C, Goodwin JS. Population-Based Assessment of Hospitalizations for Toxicity From Chemotherapy in Older Women With Breast Cancer. JCO. 2002;20(24):4636–4642. doi: 10.1200/JCO.2002.05.088 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47.Bhimani J, Wang P, Gallagher GB, et al. Patient factors and modifications to intended chemotherapy for women with Stages I–IIIA breast cancer. doi: 10.1002/ijc.35494 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 48.Bhimani J, O’Connell K, Persaud S, et al. Patient characteristics associated with delayed time to adjuvant chemotherapy among women treated for stage I–IIIA breast cancer. International Journal of Cancer. 2024;155(9):1577–1592. doi: 10.1002/ijc.35053 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 49.Bhimani J, O’Connell K, Persaud S, et al. Patient Characteristics Associated with Intended Nonguideline Chemotherapy in Women with Stage I to IIIA Breast Cancer. Cancer Epidemiol Biomarkers Prev. 2024;33(10):1286–1297. doi: 10.1158/1055-9965.EPI-24-0360 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50.Davis AC, Voelkel JL, Remmers CL, Adams JL, McGlynn EA. Comparing Kaiser Permanente Members to the General Population: Implications for Generalizability of Research. The Permanente Journal. 2023;27(2):87. doi: 10.7812/TPP/22.172 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplemental Table 1
NIHMS2149060-supplement-Supplemental_Table_1.docx (24.3KB, docx)
Supplemental Table 2
NIHMS2149060-supplement-Supplemental_Table_2.docx (21.8KB, docx)

Data Availability Statement

Data contain potentially identifiable information such as dates of events that cannot be shared openly without appropriate human subjects approval and data use agreements, but the data that support the findings of this study are available from the corresponding author upon request.

RESOURCES