Breast-conserving surgery following neoadjuvant therapy-a systematic review on surgical outcomes

In total, 22 studies mentioned margin status plus additional local therapies in patients receiving BCS after NACT [24‐45] [Table 1]. All included studies showed a low level of evidence (3 or 4) on the OCEBM scoring system. Ten comparative studies described the surgical outcomes with or without preoperative chemotherapy [24‐33]. Positive margins in these comparative studies ranged from 5 to 39.8% after NACT versus 13.1–46% without NACT, leading to secondary surgery in 0–45.4% versus 0–76.5%, respectively. Four retrospective studies reported a significantly lower number of involved margins and secondary surgery after NACT [23‐26]. One study, based on a national pathology database describing 626 patients after NACT versus 9275 patients without NACT, reported a higher rate of involved margins (27.3% vs. 16.4%) and secondary surgeries (9.1% vs. 5.3%) after NACT [33]. The other five studies showed no difference between both treatments.

Assersohn et al. reported a high rate of involved margins (39.8% vs. 36.4%) in the group receiving four cycles of neoadjuvant plus four cycles adjuvant chemo-endocrine therapy and the group receiving eight cycles of adjuvant chemo-endocrine therapy. All systemic therapy was given concomitantly with radiotherapy and their policy was not to re-resect involved margins. One patient with involved margins had an breast tumor recurrence after a follow-up of 57 months. [24] In a report by Tiezzi et al., re-excision was not performed in all 17 patients (19.3%) with close/involved margins either due to refusal or to technical impossibility (posterior margin involving the major pectoral muscle fascia). A loco-regional recurrence rate of 11% was reported [31].

Twelve non-comparing studies reported 2–33.6% of tumor-involved margins after NACT followed by BCS, with 0–12.4% receiving a re-excision and 0.7–26.5% patients receiving a secondary mastectomy [34‐45]. Two studies described secondary surgery after NACT ranging from 11.5 to 26.5%, but this was not further specified as being either a re-excision or a mastectomy [35, 41]. The presence of lobular carcinoma resulted in significantly higher degrees of tumor-involved margins compared with ductal carcinoma.[33, 36, 43] For example, Truin et al. compared lobular carcinomas with ductal carcinomas and reported a secondary mastectomy rate of 26.5% versus 4.7%, respectively [43].

Excision volumes in patients undergoing BCS after NACT were reported in fifteen studies, including two prospective cohort studies and thirteen retrospective studies [20, 25, 27, 28, 31, 33, 39, 40, 42, 44, 45, 47, 48‐50]. Thirteen studies were single-center studies. Again, all included studies showed a low level of evidence (3 or 4) on the OCEBM scoring system.

Mean excision volumes and weight ranged from 43.7–268 cm³ to 26.4–233 gr, respectively (Table 2). One study was based on data from a prospective randomized controlled trial and compared breast cancer patients receiving BCS before and after chemotherapy. Breast cancer patients with T1–T3, N0–N2 breast carcinoma were randomized between different chemotherapy regimens. The indication for administering pre- or postoperative chemotherapy was not randomized but made by the treating physicians, based on patient and tumor characteristics. In patients with T1 tumors, no significant difference was seen in excision volumes between the pre- and postoperative chemotherapy group, at 98 cm³ versus 111 cm³, respectively (p = 0.51). Tumor size at clinical presentation was 20 mm versus 15 mm (p = 0.0055). In patients with T2 or T3 tumors, the mean excision volume was significantly lower in patients who received chemotherapy preoperatively (113 cm³ vs. 213 cm³, p = 0.0043) with no difference in tumor size at presentation [25]. Four retrospective studies that compared BCS with or without NACT reported conflicting results [27, 28, 31, 33]. Two studies showed lower excision volumes after NACT (132.2 cm³ vs. 158.1 cm³, p = 0.04 and 143.6 cm³ vs. 273.9 cm³, p < 0.004) [27, 28], whereas one study showed larger excision volumes following NACT (108 cm³ vs. 78 cm³, p = 0.002) [31]. In a national pathology database study from the Netherlands, no difference was found between the use or absence of NACT in BCS (50 cm³ vs. 46 cm³, p = 0.14) [33].

Additionally, ten single-center cohort studies reported excision volumes or specimen weight after NACT without comparing these results with another group. A large heterogeneous patient population was included, and a wide range of excision volumes (43.7–268 cm³) and specimen weight (48–233 g) was reported [20, 39, 40, 42, 44, 45, 47, 48‐50].

No subanalyses regarding the predictive value of preoperative characteristics, for example, receptor status or histological subtype, and excision volumes were done in any of the fifteen studies. Four studies did analyze the influence of the response to NACT on the amount of volume excised. [33, 42, 44, 48] Volders et al. showed that the median lumpectomy volume in patients with no pathological response was 50 cc, with partial pathological response was 50 cc, and in patients with pathological (near) complete response median lumpectomy volume was 55 cc. (p = 0.018). The significant difference seems of low clinical value since there is only 5 cc difference. Moreover, patients with partial and no pathological response had significant more involved margins compared to a pathological (near) complete response (42.1, 25.6, and 12%, respectively). Ramos et al. report no difference in case of complete pathological response (24.4 gr) versus partial response (27.4 gr). [42] Peintiger et al. reported no difference in clinical complete response versus clinical partial response (75.20 cm³ vs. 66.78 cm³ p = 0.53). [48] The volume excised in patients with pCR or minimal pathological residual tumor (Payne–Miller grades 4 and 5) was significantly lower after ultrasound-guided surgery compared to wire-guided surgery. [44].

Mazouni et al. described surgical outcomes of 214 patients after BCS and 45 patients treated with oncoplastic breast surgery (OPBS) after NACT. Excision volumes were smaller in the BCS group (98 cm³) compared with OPBS (98 vs. 180 cm³, p < 0.0001)). In the BCS group, 14.5% were moderately satisfied, 47.9% were satisfied and 37.6% were very satisfied with the cosmetic outcome, which was comparable to the OPBS group (p = 0.52) [39]. Karanlik et al. described 251 patients with T2 tumors receiving BCS for invasive breast cancer between 2008 and 2011. Excision volumes were smaller in the patients receiving NACT (158.1 cm³ vs. 132.2 cm³, p = 0.04). Patients’ pictures were evaluated independently by two nurses (median 12 months after surgery) [28]. A good/excellent cosmetic outcome was more common after NACT compared to patients receiving BCS without NACT (92% vs. 80%, p = 0.03).

Based on current data, there is no evidence supporting a positive effect of NACT on tumor-free margins and consequently, on a reduction of secondary surgery. This could be due to a variety of factors. Firstly, an insufficient number of prospective, controlled studies reporting the surgical outcomes have been conducted. Secondly, preoperative imaging and estimation of, non-concentric, residual disease during preoperative imaging appears to be more difficult after tumor downstaging. As a consequence, macroscopic evaluation of the location and extent of residual disease peri-operatively is complicated. It is beyond dispute that marking of the tumor after NACT is essential to achieve identification of residual tumor or the tumor bed and clear margins after BCS. Assessment of current literature indicates that an optimal method for localizing a non-palpable lesion after NACT has not yet been established [38, 40, 42, 44].

The frequency of reported pathological complete response (pCR) after NACT has increased dramatically in the past years due to improvements in targeted therapies, with up to half of all patients in specific groups such as HER2-positive patients [51]. From the studies included in this systematic review, pCR rates range from 12 to 40.4%. [20, 31, 39, 40, 44, 47‐50] By definition, these patients will have tumor-free resection margins, and therefore, the estimated percentages of involved margins are actually an underestimate in those patients who have residual disease after NACT. Currently, while the search continues for optimal preoperative imaging to predict response to NACT, the definitive response to chemotherapy is still determined postoperatively by a pathologist.

The importance of achieving tumor-free margins in patients who received BCS after NACT is a matter of debate, especially for patients with a good response to NACT. One of the arguments is the possibility of adjuvant radiotherapy to eradicate microscopic residual tumor and thereby decrease the rate of ipsilateral breast tumor recurrence (IBTR). In the 90s, Assersohn et al. showed a low (1.0%) IBTR after 57 months in 98 patients receiving a ‘sandwich’ schedule of pre- and postoperative chemo-endocrine therapy concomitant with radiotherapy. Breast cancer stages for the patients in the NACT group were not mentioned. Unfortunately, no data are available from prospective trials with current treatment strategies regarding margin width and oncological safety following NACT. In the NSAPB-B18 trial, a significant increase of IBTR was reported in patients who were converted from mastectomy to BCS when compared with those patients who had BCS as initially planned. After 9 years, the rates of IBTR were 15.9% versus 9.9%, a difference that was no longer statistically significant after controlling for patient age and initial clinical tumor sizes [52].

The risk of bias across all included studies is high due to the retrospective character and variation among patient groups, especially in terms of selection and reporting bias. Comparison between studies is not possible and these studies therefore only reflect the current (selected) surgical outcomes after NACT. In general, the group that received preoperative chemotherapy had larger and biologically more aggressive tumors, factors that would be expected to increase the rate of positive margins and re-excision [29]. On the other hand, the studies that did not involve NACT included more lobular carcinomas, which is known to have a higher risk of involved margins.

Nonetheless, the primary goal of every surgeon when starting BCS after NACT is to excise the tumor without tumor-involved margins. Involved margins result in more frequent secondary surgery, leading to poor cosmetic outcomes, additional costs, and psychological stress for the patient. Clinical studies regarding neoadjuvant therapy in breast cancer have mainly been initiated by oncologists, resulting in a degree of neglect of surgical outcomes, a problem illustrated by the lack of data on margin status and excision volumes in randomized controlled trials.

Our view is that both oncological and surgical outcomes after NACT should be assessed equally as the latter positively contributes to a patient’s quality of life, as extensively reported in the adjuvant setting [53, 54]. Furthermore, the distress associated with secondary surgery is also underrecognized in surgical studies. For example, Truin et al. describe 30 secondary mastectomies in 466 patients with lobular cancer undergoing BCS after NACT, concluding that 6.4% of patients undergo a secondary mastectomy [43]. The correct calculation would have included primary lumpectomies only, resulting in a figure of 26.5% (30 secondary mastectomies of 113 primary lumpectomies) (Fig. 1). Patients with invasive lobular carcinoma will have lower pathological complete response (pCR) rates and higher rates of secondary surgery due to positive margins and should therefore be informed about these negative outcomes.

Neoadjuvant chemotherapy could potentially improve cosmetic outcomes by reducing the volume of excised breast tissue in candidates for both mastectomy and BCS. Among patients with T2 tumors, two studies showed that patients treated with NACT underwent less extensive excision compared with those who underwent primary surgery. Moreover, the smaller resected volumes did not lead to increased risk of re-excision to obtain negative margins [25, 28]. By contrast, Tiezzi et al. reported larger resection volumes in patients after NACT, although it should be noted that these tumors were larger at presentation [31]. It is important to realize that larger volumes do not necessarily result in fewer involved margins or less additional therapy.[16, 25, 27, 47].

With the advent of NACT, the challenge for surgeons became to localize the remaining lesion and to resect the minimum amount of healthy breast tissue, while achieving tumor-free margins.

The pattern of tumor regression is not always concentric and may also occur with diffuse fragmentation. Therefore, the concern of leaving microscopic residual tumor surrounding the surgical area after NACT resulted in a dilemma of how much breast tissue to excise. Plainly, it is not necessary to excise the original tumor volume, otherwise all advantages of performing NACT for downstaging of the breast will vanish [25].

Due to the retrospective and non-comparative design of the currently available single-center cohort studies, no general conclusion can be drawn. The results only reflect the current volumes resected after NACT. Cosmetic failure rates have proven to be significantly higher if the size of the lumpectomy exceeds 40–100 cc, regardless of breast size [9, 12, 13, 18]. Keeping this in mind, almost all of the studies report extremely high excision volumes after NACT, probably resulting in poor cosmetic outcomes for these patients. It would have been of great value if certain subgroups were analyzed for their predictive value on resection volumes, but unfortunately no conclusion can be drawn from the current literature regarding the impact of, for example, receptor status, histological subtype, or chemotherapy regimen on excision volume and cosmetic outcome. The calculated resection ratio (CRR), which is the total resection volume divided by the optimal resection volume, is a measure to assess excessive breast tissue resection. Only one study reported a resection ratio, with a median CRR in the primary surgery and neoadjuvant therapy groups of 3.3 and 2.0, respectively (p < 0.0001). This implies that the excision volumes were 2–3.3 times as large as they should be [55].

With improvements in NACT strategies, and the increasing rates of pCR with a good prognosis, cosmetic outcomes are becoming increasingly important to breast cancer patients. Unfortunately, any evidence supporting an improvement of cosmetic outcomes after tumor downstaging with NACT is currently lacking. Two single-center retrospective studies reported acceptable results, but included small and highly selected study groups, with a large risk of bias [28, 39]. A poor cosmetic outcome of the breast following breast cancer treatment has a high impact on patients’ quality of life, being a daily reminder of their previous breast cancer and of their treatment period. For this reason, a thorough counseling of patients on the expected primary and secondary outcomes of their treatment is mandatory. In particular, patients receiving NACT to convert from mastectomy to BCS, may have high expectations regarding their cosmetic outcomes and may therefore be unprepared for poor cosmetic results. In particular, when realizing that cosmetic outcomes after a mastectomy with a breast reconstruction may have very satisfactory results as well [7].

One of the main indications for NACT in breast cancer patients to date is tumor downstaging to achieve less morbidity and improved cosmetic outcomes. However, it is interesting that these outcomes of NACT are not supported by any scientific evidence.

The motivation for use of NACT to downstage the breast tumor is based on two facts regarding BCS without NACT: However, these statements are both based on studies without NACT. Although BCS appears technically feasible in patients receiving NACT after tumor downstaging, a satisfactory cosmetic outcome cannot be guaranteed when still a large amount of breast tissue is resected.

The negative influence of large resection volumes on cosmetic outcomes and the frequent poor cosmetic outcome of BCS without NACT validate current efforts to improve cosmetic outcomes with oncoplastic breast surgery (OPBS). The theoretical advantage of oncoplastic surgery for breast cancer is the possibility of using wider resection margins, leading to improved oncological outcomes and less secondary local therapies, combined with a good cosmetic result. However, involved margins and secondary local therapies after OPBS are still frequently reported [56, 57]. Unfortunately, current literature on OPBS without NACT largely consists of poorly designed and underpowered studies, while high-quality literature on OPBS in patients receiving BCS after NACT is absent. Mazouni et al. reported high rates of involved margins (15.6%) and secondary mastectomies (25%) in OPBS after NACT, while no difference in cosmetic outcome was seen compared to BCS after NACT.[39] Amabile et al. reported secondary mastectomy in 27.3% of patients after NACT followed by OPBS [32].

In conclusion, the advantages of NACT in terms of lower resection volumes and improved cosmetic outcomes after BCT have not yet been proven. Prospective randomized trials including secondary local therapies, resection volumes, and cosmetic outcomes in patients receiving postoperative chemotherapy, as well as reporting the type of initial surgery, will be required to confirm the hypothesis that NACT improves cosmetic outcomes by lowering excision volumes. In a new era of neoadjuvant treatments, it is imperative for surgeons to include not only outcomes of surgical techniques in their routine clinical practice, but also to take note of patient outcome measures in order to better monitor and compare surgical treatment outcomes in meaningful ways.