China in December 2019 1 , and the World Health Organization (WHO) declared a pandemic in March 2020. 2 By mid-2022, over 536 million confirmed cases and more than 6.3 million deaths were reported globally. 3 This has had a huge impact on the management of resources towards the care of COVID-19 patients. In addition to this, government enactments of stay-at-home orders had a direct effect on the care of chronic disease and cancer patients.

Before the pandemic, nearly two and a half million women around the world were diagnosed with breast cancer each year. 4 The pandemic has potentially had an effect on breast cancer diagnosis and treatment for these millions of women due to changes in healthseeking behavior and reduced availability of screening and treatment services. 5 Understanding the changes that occurred due to the pandemic is essential as women may be at risk of poor cancer outcomes due to suboptimal treatment. 6 Reviewing the experience in cancer treatment centers may also assist in planning for future variants, pandemics or other major disruptions to health care. This study has systematically reviewed the published literature and explored changes to breast cancer treatments (surgery, radiotherapy, chemotherapy and endocrine therapy) during the COVID-19 pandemic by comparing 'pandemic' treatment to 'pre-pandemic' or 'ideal' treatment.

This review was registered on PROSPERO (Study ID CRD42021279655) 7 and it was performed using PRISMA methodology. 8

Studies were eligible if they reported the management of breast cancer during the early phase of the COVID-19 pandemic (January 2020-June 2021) and used a paired design, which included a comparison group (either a non-COVID time period or an individual patient comparison to 'usual' or 'ideal' treatment that would have been delivered during a non-COVID period). Studies reporting surgery, chemotherapy, radiotherapy and/or primary endocrine therapy were eligible. A change to treatment was defined as a delay, replacement of one treatment with another, change in the sequencing of treatment modalities or omission of one or more standard treatment options.

Additional eligibility criteria were: original studies, published in full in English in the peerreviewed literature. Exclusion criteria were abstractonly publications, letters, review articles and studies reporting treatment during the pandemic that did not include a comparison group.

A search of EBSCO (including CINAHL, MEDLINE and Psych INFO), PubMed and Cochrane Library was performed using the search terms:

("Breast Cancer" OR "Breast neoplasm") AND (Surgery OR Chemotherapy OR Radiotherapy OR Treatment OR Management) AND (Pandemic OR COVID OR "COVID-19" OR coronavirus OR "SARS COV-2")

Following removal of duplicates, titles and abstracts were screened for eligibility by both authors.

Full-text papers were retrieved and re-screened by both authors.

Data were extracted by one author (MSB) and checked by the other (MEB). Data were collected on study characteristics (study setting, design, population, comparison group, number of participants), COVID setting (country, associated COVID protocols/ health order, COVID prevalence in study group) and cancer treatment outcomes (presentation, change or delay to treatment types). Data were extracted into a spreadsheet. Data were analyzed by grouping studies according to outcomes of interest and treatment type and comparing the results. Results were presented in narrative form.

The Newcastle-Ottawa Quality Assessment Form for Cohort Studies was used to assess the risk of bias. It measures the quality of nonrandomized studies based on three criteria: selection of the study groups; comparability of the groups; and the ascertainment of either the exposure or outcome of interest. It produces a score for each criterion with a total maximum score of 7. 9 Risk of bias was independently assessed by each author and consensus was reached after discussing discordant results.

The initial search identified 485 abstracts. After screening, 15 full-text studies met our eligibility criteria 101112131415161718192021222324 (Figure 1: PRISMA flowchart). These studies reported outcomes for 6,353 people treated for early breast cancer from January 2020 to June 2021.

Study characteristics are shown in Table 1. There were eight studies from Europe, 1012131419212324 five from North America, 1115161722 and two from Asia. 1820 Eleven studies compared treatment during the COVID period to a similar non-COVID period. 1011131415171920212324 There were four, 12161821 prospective and 11 retrospective studies. 1011131415171920222324 Four compared the actual treatment of individual patients during the COVID to the 'usual' or 'ideal' treatment that would usually have been delivered in a non-COVID period. 12161822 For treatment modalities, 14 studies included data for surgery 1011121314151618192021222324 , ten for neoadjuvant chemotherapy, 10121314151618212324 four for adjuvant chemotherapy, 12182224 nine for neoadjuvant endocrine therapy, 101213151618212224 seven for adjuvant radiation therapy, 10121517182224 and four studies included data on a 'no treatment' group. 12131618 The studies were all conducted in settings that were negatively affected by the COVID-19 pandemic. Specifically, 10 studies were under lockdown conditions, which included stay-at-home orders. 10121416171821222324 Five studies were conducted at times when services were affected by challenges to the hospital or health system due to COVID-19 infection in the community. 1113151920 Risk of bias Assessment using the Newcastle-Ottawa Quality Assessment Form for Cohort Studies 9 indicated a very low risk of bias across the studies (Table 1). Thirteen studies received the maximum scores of 4, 2, and 1 for the three individual components (total score of 7, indicating low risk of bias). 10111213141516171819222324 Two studies received scores of 4, 1 and 1 (total score of 6). 2021 Both studies scoring 6 lost a point under the Outcome/Exposure domain due to not reporting follow-up data on the cohort. This was not considered to be a significant bias for this review, which is focused on initial treatment of breast cancer rather than longer-term outcomes. Table 2 shows the characteristics of cancer presentations. Six studies reported a significant decrease in the number of cases treated during COVID-19 period. 101315172023 The difference ranged from a 19-33% decrease in cases compared to non-COVID-19 period. 1315 Three studies reported an increase in the number of cases: one reported a non-significant increase, 11 and two others reported a 7% and 18% increase. 1924 One of these studies attributed the increase to receiving cancer patients from designated COVID hospitals in the region, 19 while the other study attributed the increase to the exclusion of nononcological breast surgery cases (mostly breast reconstruction), which were more common in the pre-COVID control group. 24 The remaining 5 studies did not report on the difference in the number of cases during COVID-19 period. 1216182122 *likely overlap in patient populations in the two Vanni studies; **Quality/risk of bias score Newcastle-Ottawa Quality Assessment for Cohort Studies nr=not reported; na=not applicable; BCS=breast conservation surgery; BR=breast reconstruction; ET=endocrine therapy; WLE=wide local excision; SLNB=sentinel lymph node biopsy; ALND=axillary lymph node dissection; *likely overlap in patient populations in the two Vanni studies

Most studies reported no difference in the age of the cohort during the COVID period (Table 2). 10111415202324 Two studies reported a younger COVID cohort compared to non-COVID cohort and. 1217 Two studies reported that older patients are more likely to delay surgery or have a change to their treatment. 1822 Six studies reported a higher proportion of invasive cancers (compared to DCIS cases) and a higher proportion of more advanced tumour stage in the COVID cohort. 151920212324 Three studies reported that the group with higher tumour stage were less likely to experience delays of changes to their treatment. 161822 Two studies reported no difference in tumour characteristics between the COVID and control group. 1314 Changes to treatment modalities Surgery Changes to surgical treatment are shown in Table 3. Several aspects of surgical treatment were changed during the COVID-19 period. For surgical procedure, six studies reported no difference to the type of breast surgery during the COVID-19 period. 101113162024 This included comparisons between breast conserving surgery (BCS), mastectomy, breast reconstruction (BR). Three studies reported an increase in cases and proportion of surgical cases undergoing BCS. 182123 The same studies also reported a decrease in mastectomy with BR. One study reported a decrease in the proportion of BCS and BR with an increase in simple mastectomy (without reconstruction). 14 One study reported a decrease in prophylactic surgery and increase in nipple-sparing mastectomy. 19 Some of the studies also reported data on axillary surgery during the COVID-19 period. These include data on axillary lymph node dissection (ALND) and sentinel lymph node biopsy (SLNB). Four studies reported an increase in ALND proportion, along with a decrease in SLNB. Three studies reported no differences in the proportion or number of ALND and SLNB procedures.

Most of the studies reported whether changes to waiting time for surgery/treatment occurred. Five studies reported no difference in the wait time during the COVID-19 period. 1011141520 Six studies reported an increased wait time during the COVID-19 period. 12131618212224 The additional wait time for surgery (delay) ranged from 8 to 47 days. No study reported a shorter wait time for treatment during the COVID-19 period. One study grouped surgical delay into cases that were directly related to service issues, where delay was unavoidable, and cases where patients elected to delay their surgery to avoid exposure to the virus and/or additional stress on the system. 18 It is unclear in the remaining studies how much of the reported delay may have been due to 'system' issues versus patient choice.

Five studies reported a shorter hospital stay after surgery. 1011141924 Some studies showed an increase in proportion of same day discharges, 1119 others show a 0.7 to 1-day decrease in hospital stay during the COVID-19 period. 1014 One study reported no difference in hospital stay duration. 24 Some of the studies also noted additional data relating to changes in surgery during the COVID-19 period. One study reported an increase in regional anaesthesia (relative to general anaesthesia), and another study reported a decrease in the use of regional anaesthesia. One study reported an increase in telehealth usage for surgical follow-up during the COVID-19 period. Two studies reported no difference in re-operation rate, readmission and surgical complications.

Chemotherapy Three studies reported no change in the use of neoadjuvant chemotherapy (NACT) during the COVID-19 period. Three studies reported a decrease in the use of NACT; one specified the decrease was in Stage I and II patients. Another study noted the substitution of NACT with endocrine therapy instead. Two studies reported an increase in the use of NACT during the COVID-19 period.

Three studies reported using primary systemic therapy (PST, data for NACT and NAET reported together). One had an increase in its use, 22 and two others reported no changes. 1016 Three studies reported a decrease in or omission of adjuvant chemotherapy (CT), while one reported no significant difference in CT cases during the COVID-19 period.

Radiotherapy Six studies reported data for radiotherapy. 121517182223 In one study, fewer women commenced treatment in the study period, 17 and in another there was an increase (in a study that had more BCS cases during the pandemic). 23 Another study reported changes to treatment regimens, with an increase in recommendations for omission or hypofractionation (15 fractions reduced to five). 12 Endocrine therapy Studies reported increased use of neoadjuvant endocrine therapy. Four studies reported increased use of neoadjuvant endocrine therapy (NAET) 12131516 . These include a general increase in the use of NAET to postpone non-urgent surgery and as a replacement for neo-adjuvant chemotherapy. One study reported no significant difference in the use of NAET during the COVID-19 period, 23 and the remaining studies did not report outcomes for NAET.

This systematic review evaluated the impact of COVID-19 on breast cancer treatment. Fifteen studies with a paired design were identified and these were all of high quality (low risk of bias) despite the challenges of conducting research during the pandemic.

The extent of the impact of the pandemic varied across studies. Most studies reported stable numbers or fewer cases of breast cancer presenting for treatment during the pandemic. A decrease is unsurprising given the lockdowns and restrictions in access to health care, including screening and diagnostic services. However, many studies reported an increase in the proportion of later stage cancers, suggesting that the 'missing' cancers (predicted but not diagnosed) are likely to be early cancers that would have been detected by screening during this period. One hypothesis could be that mortality from breast cancer will not be impacted by the pandemic if these early-stage cancers are identified in the early months of COVID recovery (while delayed but still early-stage). The increase in case numbers was seen in one study appeared to be due to restructuring of local health services for dedicated COVID-19 treatment hospitals rather than a true increase in case numbers in the health jurisdiction. 19 Therefore, the numbers in individual hospitals may not always reflect the total number of cases in a health district or state/province and population-level data is required to examine the true incidence of breast cancer during 2020 and 2021.

There was a general trend towards increased wait time for treatment or increased delays during the COVID-19 pandemic. 12131618212224 There was also a trend towards reduced hospital stay duration and increased number of same-day discharges. 10111419 This is likely to represent an aim to minimise the amount of time that patients were in hospital to reduce the risk of cancer patients contracting COVID-19 and to minimise the burden of non-COVID cases in the hospital. Services demonstrated an ability to adapt to the local conditions by performing more breast conservation surgery or using more neoadjuvant chemotherapy and endocrine therapy in response to the challenges of admitting patients to hospital for surgery. In some studies, delay to surgical treatment was intentional and even chosen by patients in consultation with their doctors. In one study, women who were older, with ER-positive early-stage breast cancer were more likely to take up the option of delayed surgery combined with neoadjuvant endocrine therapy. 12 Most of the studies reviewed in this paper reported on surgical therapy. Other treatment modalities were less frequently reported. While most studies reported no change in the type of surgical procedure performed, some reported an increase in breast conserving surgery (including a study that reported an increase in oncoplastic procedures). 21 The explanation for this is unclear. It may reflect an attempt to conserve the breast in women who may have otherwise undergone mastectomy, to minimise hospital stay and avoid breast reconstruction surgery, which was suspended by many centres during the pandemic. The increase in the proportion of breast conservation procedures may also be related to the fact that neoadjuvant chemotherapy or neoadjuvant chemotherapy was used to delay surgery in some cases, resulting in tumour shrinkage. However, the increase in breast conservation could be expected to result in an increasing proportion of women requiring breast radiotherapy. Only six studies discussed the impact of COVID on breast radiotherapy and these generally indicated a trend towards hypofractionation and partial breast irradiation protocols rather than an increase in radiotherapy case numbers overall. An increase in axillary lymph node dissection with a decrease in sentinel lymph node biopsy was observed in some studies. 10212324 This may imply more advanced stage of cancer in surgical patients, and it is difficult to reconcile with the increase in breast conservation.

It was rare for centres to continue performing breast reconstruction during the pandemic. 19 While some patients with more extensive in-breast disease may have been treated with extended oncoplastic conservation procedures, 20 others may have been forced to accept simple mastectomy without reconstruction. 161824 This is consistent with other cohort studies during the pandemic that indicated breast reconstruction was unavailable in many places. 1012141825 Lack of access to immediate breast reconstruction may be one of the long-term negative quality-of-life impacts of cancer during the pandemic. Long-term follow-up of these cases is needed to investigate the number of women who access delayed reconstruction procedures and the impact of waiting for reconstruction or not having the option at all.

This study has some limitations. The included studies are heterogenous in the method they used to report changes in treatment due to COVID-19, so pooling the results was not possible. Publication bias was not assessed, and this may have had an impact on the studies that were available for inclusion.

The 15 studies included in this review demonstrated a significant impact of COVID-19 on breast cancer treatment. The changes to treatment were not consistent across the studies. This reflects the individual solutions that cancer centres found, and these were determined by their local circumstances. Delaying surgery by using more neoadjuvant therapy was seen in six of the 15 studies. 121518202123 A preference for simpler surgical procedures (breast conservation or simple mastectomy) was seen and most centres were unable to provide immediate breast reconstruction. A higher proportion of later-stage cancers was seen, suggesting that more serious cancers presented for treatment and that early screendetected cancers were not identified during this period. Ongoing observation of the cohort of women treated during the pandemic, and correlation with population-level incidence and survival data is essential to fully understand the long-term impact of COVID-19 on breast cancer. Lack of access to breast reconstruction may have a lasting negative effect on quality of life and this should also be explored with follow-up studies.