Management of ipsilateral breast tumor recurrence following breast conservation surgery: a comparative study of re-conservation vs mastectomy

Abstract

Background:

Breast conservation therapy (BCT) is well-established for the management of primary operable breast cancer, with oncologic outcomes comparable to those of mastectomy. It remains unclear whether re-conservation therapy (RCT) is suitable for those patients who develop ipsilateral breast tumor recurrence (IBTR), for whom mastectomy is generally recommended.

Methods:

We identified women who underwent BCT for invasive or ductal carcinoma in situ and developed IBTR as a first event, comparing the pattern of subsequent events and survival for those treated by RCT versus mastectomy.

Results:

Of 16,968 patents who had BCT, 322 (1.9%) developed an isolated IBTR as a first event between 1999–2019. 130 (40%) had RCT and 192 (60%) mastectomy. Compared to mastectomy, the RCT patients were older (66 vs 53, <0.001), had a longer disease-free interval (DFI: 5.8 vs 2.7 years (p<0.001)), were less likely to have received RT (p<0.001), endocrine therapy (ET) (p<0.005) or combined RT/ET (<0.001) as initial treatment, but the characteristics of their initial primary cancers and of their IBTR were comparable. At a median follow up of 10.7 years following initial BCT and 6.5 years following IBTR, there were no differences in BCSS or OS between RCT and mastectomy.

Conclusion:

For BCT patients who developed IBTR as a first event, we observed comparable BCSS and OS from time of initial treatment and from time of IBTR, whether treated by RCT or mastectomy. These results support wider consideration of RCT in the management of IBTR, especially in the setting of older age and longer DFI.

Mini-abstract

For ipsilateral breast tumor recurrence (IBTR) following breast conservation therapy, we observed comparable overall and breast cancer-specific survival whether treated by re-conservation (RCT) or mastectomy. These results support wider consideration of RCT in the management of IBTR, especially in the setting of older age and longer disease-free interval.

Introduction

Breast conservation therapy (BCT) has become standard care for women with early-stage breast cancer, with 7 randomized trials demonstrating comparable oncologic outcomes for BCT vs mastectomy at 10 years, including equally low rates of local recurrence.¹ In contrast, the management of ipsilateral breast tumor recurrence (IBTR) remains a subject of debate, with current National Comprehensive Cancer Network (NCCN) guidelines continuing to recommend mastectomy² – the historic norm - as treatment for IBTR, driven in part by concerns about re-irradiation (RT), yet without clear proof of a survival benefit. A number of prior studies demonstrate – with significant limitations - the feasibility of re-conservation therapy (RCT) with good outcomes in selected patients.^3–9 Here we aim to compare in detail the characteristics and oncologic outcomes in our own BCT patients with isolated IBTR treated by RCT vs mastectomy.

Materials and Methods

Under a Waiver of Authorization from the Memorial Sloan Kettering Cancer Center (MSKCC) institutional Review Board (IRB), we searched our prospectively maintained surgical database for patients initially treated with BCT and presenting between 1999 and 2019 with isolated IBTR as a first event. We excluded patients with contralateral breast cancer, BRCA mutations, and concurrent regional node or distant metastasis. Our standard post-BCT follow up comprised physical examination 1–2 times a year, annual mammography, and annual ultrasound (for women with dense breasts). MRI was not routine, but was done selectively for problem-solving. For those subsequently treated by RCT vs mastectomy we compare in detail tumor characteristics, extent of initial surgery, use of systemic therapy and RT, disease-free interval (DFI), features of the IBTR, pattern of subsequent events (local, regional and distant), and both breast cancer specific (BCSS) and overall survival (OS). The crude median duration of follow-up for all patients from time of initial treatment was 10.7 years and from time of IBTR was 5.6 years for RCT vs 7.1 years for mastectomy.

We used appropriate t tests (excluding unknown values) to compare frequencies and the Kaplan-Meier method to determine BCSS and OS. All statistical analyses were conducted using R 3.5.2 (R Core Team 2017, Vienna, Austria), setting the type I error rate at 0.05α.

Results

Of 16,968 patients who underwent BCT (1986–2018), 1,912 (11.2%) presented with cancer recurrence (1999–2019), 413 (2.4%) of these were locoregional recurrences, among which 322 (1.9%) were isolated IBTR as first event. Of these, 130 (40%) had RCT and 192 (60%) mastectomy. Among 116 (36%) who had their initial BCT outside our institution, all had review of outside pathology slides at our institution.

Among patients with isolated IBTR there was a significant trend toward more use of RCT in recent years, increasing from 27% (2000–2004) to 61% (2015–2019) (Figure 1, p<0.001).

Figure 1:

Increasing frequency of RCT over time

Although the initial tumor characteristics for RCT vs mastectomy were roughly comparable (Table I), patients treated by mastectomy were younger (50 vs 58, P<0.001), with a higher proportion of invasive (86% vs 75%, p=0.041) and triple negative disease (23% vs 11%. P=0.014).

Table 1:

Clinicopathological characteristics of initial cancer

Characteristic	RCT n= 130 # (%)	Mastectomy n=192 # (%)	p-Value
Median age, years (IQR)	58 (49,68)	50 (44,58)	<0.001
Median size invasive tumor, cm (IQR)	1.2 (0.8, 1.8)	1.2 (0.7, 2.0)	0.88
Tumor Size			0.11
In situ	32 (25)	27 (14)
≤2 cm	77 (59)	121 (63)
>2 cm	20 (15)	39 (20)
Unknown	1 (1)	5 (3)
HistologyΔ			0.041
DCIS	32 (25)	27 (14)
IDC*	90 (69)	148 (77)
ILC	6 (5)	14 (7)
Other	2 (2)	3 (2)
Regional nodal involvementƥ			0.62
pNx	30 (23)	26 (14)
pN0	78 (60)	121 (63)
pN1	16 (12)	35 (18)
pN2	4 (3)	9 (5)
pN3	2 (2)	1 (1)
Hormone Receptor Subtype Invasive Cancer
ER+/HER2−	68 (52)	83 (43)
HER2+	13 (10)	35 (18)
TNBC	14 (11)	44 (23)	0.014
Unknown	3 (2)	3 (2)
Focality			0.11
Multifocal	27 (21)	56 (29)
Presence of LVI	28 (22)	44 (23)	0.53
Hormone Receptor Status for DCIS
ER+	6 (5)	7 (4)
ER−	1 (1)	1 (1)
Unknown	25 (19)	19 (10)
Time of initial operation
Prior to 2005	44 (34)	66 (34)
After 2005	86 (66)	126 (66)

RCT, re-conservation therapy. IDC, Invasive Ductal Carcinoma. ILC, Invasive Lobular Carcinoma, DCIS, Ductal Carcinoma in situ IQR, interquartile range; HR+/HER2−, Hormone Receptor and HER2 negative; LVI, Lymphovascular Invasion; TNBC, triple negative breast cancer; pNx: nodal stage not assessed. pN0: No positive nodes on final specimen; MSKCC: Memorial Sloan Kettering Cancer Center; OSH, Outside Hospital. ER+: Estrogen Receptor Positive.

^*includes 20 patients with ductal carcinoma in situ with microinvasion (4 in RCT and 16 in TM)

^ƥ:p value calculated based on pN0 vs pN+

^Δ:p value calculated based on invasive vs non-invasive recurrences

Compared to mastectomy (Table 2), patients who had RCT were less likely to have undergone RT (66% vs 90%, p<0.001), chemotherapy (34% vs 51%, p = 0.024), endocrine therapy (ET: 40% vs 48% p = 0.005), or combined RT/ET (34% vs 47%, p<0.001) at initial treatment. Thirty-nine (12%) patients did not receive any adjuvant therapy, 30 (23%) in the RCT and 9 (5%) in the TM group.

Table 2:

Adjuvant therapy for initial cancer

Treatment	Total # (%)	RCT n= 130 # (%)	DFI (yrs)	Mastectomy n=192 # (%)	DFI (yrs)	p-Value
RT	258 (80)	86 (66)	7.5	172 (90)	3.3	<0.001
CT*	142 (44)	44 (34)	7.1	98 (51)	3.3	0.024
ETŦ	144 (45)	52 (40)	7.4	92 (48)	3.4	0.005
RT + ET	135 (42)	44 (34)	8.3	91 (47)	3.5	<0.001
Combination Therapy for Invasive Cancer
CT + ET	79 (25)	27 (21)	7.8	52 (27)	3.4	0.55
RT+ CT + ET	73 (23)	25 (19)	8.2	48 (25)	3.5	0.59

RT: radiation Therapy; CT: chemotherapy; ET: endocrine therapy

^*calculated only for invasive disease (DCIS not included)

^Ŧ:Triple negative breast cancer and ER negative DCIS not included in this analysis

Compared to mastectomy (Table 3), patients having RCT had a significantly longer DFI (5.8 vs 2.7 years, p<0.001), were older (66 vs 53, p<0.001), were more likely to have “other quadrant” rather than “initial site” recurrence (38% vs 19%, p=0.005), and were less likely to have axillary surgery for their IBTR (23% vs 65%, p<0.001). Among all patients with isolated IBTR only 10 (3%) patients had axillary node disease. Interestingly, for RCT vs mastectomy, 12 (9%) vs 31 (16%) of patients had no residual cancer in their final surgical specimens (i.e. all disease had been removed by the core biopsy procedure).

Table 3:

Characteristics of IBTR – RCT vs mastectomy

Characteristic	RCT n= 130 # (%)	Mastectomy n=192 # (%)	p-Value
Median age (IQR)	66 (55,72)	53 (47,62)	<0.001
Median DFI - years (IQR)	5.8 (1.7, 7.4)	2.7 (1.6,4.1)	<0.001
IBTR detected by			0.027
Mammogram	84 (65)	106 (55)
Palpation	14 (11)	39 (20)
MRI	16 (12)	31 (16)
Ultrasound	12 (9)	11 (6)
Other/ Unknown	4 (3)	5 (3)
Median size IBTR - cm (IQR)	1.1 (0.6,1.4)	1.0 (0.4,1.8)	>0.99
In situ	29 (22)	33 (17)
≤2cm	74 (57)	101 (53)
>2cm	13 (10)	22 (11)
No residual	12 (9)	31 (16)
Unk	2 (2)	5 (3)
Node status			<0.001
pNx	85 (65)	45 (23)
pN0	42 (32)	135* (70)
pN1	2 (2)	7 (4)
pN2	1 (1)	1 (1)
Unk	0 (0)	4 (2)
Histology			0.83
DCIS	36 (27)	48 (25)
IDC	79 (61)	126Ŧ (66)
ILC	11 (8)	13 (7)
Other	4 (3)	5 (3)
Location (quadrant) of IBTR			0.005
Same	73 (56)	117 (61)
Different	49 (38)	36 (19)
Unk	8 (6)	39 (20)
Receptor Subtype of IBTR			0.49
Same	59 (45)	88 (46)
Different	29 (22)	52 (27)
Unk	42 (32)	52 (27)
Status at last followup			0.48
Alive	113 (87)	173 (90)
Died	17 (13)	19 (10)

^*includes three with micrometastasis in upfront surgery (no neoadjuvant chemotherapy administered)

^Ŧ:includes 9 patients with DCIS-M

RCT Repeat conservation therapy, MMG Mammogram, MRI magnetic resonance imaging, pNx, no axillary surgery performed; pN0, no positive nodes on final specimen; pN1, 1–3 positive nodes on final pathology; pN2, 4–9 positive nodes on final pathology; DCIS, Ductal Carcinoma in situ; IDC, Invasive Ductal Carcinoma; ILC, Invasive Lobular Carcinoma; IBTR, Ipsilateral Breast Tumor Recurrence.

Compared to those undergoing mastectomy (Table 4), patients having RCT were less likely to receive chemo-, ET, and chemo-ET, but more likely to receive RT following resection of the recurrence (59% vs 11%, p<0.001). Among the 130 RCT patients, 41 (32%) received re-RT, of whom 4 (10%) had another locoregional event. Among all 64 patients with IBTR who did not receive RT at initial treatment, 44 (69%) had RCT.

Table 4:

Adjuvant Therapy for Recurrence

Adjuvant Treatment for Recurrence	RCT n= 130 # (%)	Mastectomy n=192 # (%)	p-Value
CT	33 (25)	74 (39)	0.005
ET	44 (34)	82 (43)	0.047
RT	77 (59)	21 (11)	<0.001
CT + ET	10 (8)	36 (19)	0.005

CT: chemotherapy, ET: endocrine therapy, RT: radiation Therapy

Following treatment for IBTR (Table 5), there were further events in 17% of RCT vs 15% of mastectomy patients (p<0.001), with the greatest number in the ipsilateral breast following RCT and at distant sites following mastectomy, and with few regional node recurrences or contralateral cancers. For previously irradiated BCT patients with invasive IBTR treated with RCT, there were more subsequent breast events without re-RT (5 of 29 (17%)) than with re-RT (1 of 33 (3%)), but no differences in regional node (0 vs 1), contralateral (1 vs 0) or distant events (1 vs 0). In a multivariate analysis incorporating the initial tumor and treatment characteristics (age, T/N/ER status, LVI, systemic therapy (chemo/hormonal), and RT, plus DFI and operation for IBTR, there were no significant associations with BCSS.

Table 5:

Third Events after IBTR

Event	RCT n=130 # (%)	Mastectomy n=192 # (%)	p-Value
Alive and Event Free	91 (70)	147 (77)
Re-recurrence	22 (17)	28 (15)	< 0.001
DFI (to re-recurrence) - years	3.4	3.8
Local recurrence	17* (13)	4 (2)
Regional recurrence	3 (2)	8Δ (4)
Contralateral breast cancer	2 (2)	3 (2)
Distant disease	0 (0)	13 (7)
Deaths			0.09
Breast cancer	3 (2)	9 (5)
Other causes	14 (11)	10 (5)

RCT: Repeat conservation therapy. IBTR: Ipsilateral breast tumor recurrence. DFI: Disease Free Interval

^*One patient with presented with IBTR and simultaneous contralateral breast cancer

^Δ:Two patients initially presented with regional recurrence, but were found to have distant disease within 3–9 months of diagnosis

Whether calculated from time of initial BCT or from time of surgery for IBTR, there were no significant differences in BCSS or OS between RCT and mastectomy for all patients (Figures 2A–D), or when stratified by age <60 vs ≥ 60 at initial BCT (BCSS p=0.087 vs 0.26 and OS p=0.66 vs 0.54, respectively).

Figure 2A-D:

OS and BCSS from time of initial treatment and time of IBTR

A: OS from Initial Cancer Operation, B: BCSS from Initial Cancer Operation

C: OS from Operation for IBTR, D: BCSS from Operation for IBTR

Discussion

BCT is well established as oncologically equivalent to mastectomy, with comparably low rates of local recurrence in randomized trials (6.2% vs 5.9% at 10 years)¹ but for BCT patients who develop IBTR, mastectomy remains the norm and is supported by current NCCN guidelines² despite a lack of evidence to support a survival benefit.

There are a number of reasons to question this dogma, many supported by our own data. With widespread screening, as many as 17% of new breast cancer diagnoses are DCIS,¹⁰ a condition for which BCT is usually suitable, RT need not be routine, and the risk of distant relapse is negligible, either at the outset or at the time of IBTR.^11,12 With advances in breast imaging, isolated IBTR are typically detected as small non-palpable lesions (in our series, 1.1 cm in size, 90% non-palpable, and no residual cancer found in 9–16% of surgical specimens), most of these suitable for RCT relative to the size of the breast. Preoperative localization of non-palpable lesions and processing of surgical margins¹³ have both become more precise in recent decades, resulting in lower rates of re-excision and better prospects for good cosmesis following BCT and RCT. With advances in systemic therapy (chemo- and hormonal) local recurrence represents a declining proportion of all first events post-treatment.¹⁴ It is also clear that survival following IBTR is sufficiently poor (5 year OS was 76% post-IBTR in 5 NSABP trials of BCT for node-negative and 60% in 5 trials for node-positive disease)^15–16 that systemic therapy makes sense for all invasive IBTR.¹⁷ In this setting, prolonged hormonal therapy and newer agents targeting subtype will play a growing role. Finally, the options for RT, both as initial treatment and for IBTR, are increasingly nuanced. Many women with DCIS^11,12 and a growing proportion of those over age 70 with early stage in invasive disease do not require RT as initial treatment¹⁸, leaving open the option for RT following RCT for IBTR. Absent prior RT, adjuvant radiotherapy becomes a straightforward proposition after RCT. Moreover, re-irradiation (typically partial breast) is the subject of many small retrospective studies and – more recently – of prospective trials demonstrating feasibility in selected patients.^8,19,20 Our experience with re-irradiation is the subject of a separate report (in preparation)²⁰ but since 2011 has followed the protocol of the RTOG 1014 trial (partial breast, 45 Gy in 30 1.5 Gy fractions given twice a day)⁸. In short, for IBTR one size no longer fits all.

The concept of RCT for IBTR is not a new one. In 1991, Kurtz et.al.³ reported the results of RCT in 50 patients with IBTR; subsequent 5 year local control on multivariate analysis was significantly related only to DFI (92% vs 48% for DFI >5 vs ≤5 years) and margins of the RCT specimen (73% vs 36% for negative vs positive margins). In 2005, Alpert et.al.⁴ compared the results of RCT (30 patients) vs mastectomy (116 patients), observing comparable survivals on 10-year follow-up: 58% vs 66%. In 2007 and 2012, Gentilini et.al.^5–6 reported results of RCT in 161 patients; 5-year OS was 84% and the 5-year rate of further local recurrences ranged from 15% (for IBTR ≤2 cm and DFI >48 months) to 71% (for IBTR >2 cm and DFI ≤48 months). In 2011, Ischitobi et.al.⁷ reported RCT in 78 patients; local control at 5 years was 79%, and far better for ER+ than for ER- tumors: 88% vs 55%. In 2019 Sallem et.al.⁹ compared RCT (47 patients) with mastectomy (74 patients) and at 14 year follow up noted comparable local control, 83% vs 86%, respectively, but with a major difference in DFI between groups, 12 years for RCT vs 7 for mastectomy. Taken together, these studies and others indicate the feasibility of RCT with excellent outcomes given case selection on the basis of longer DFI, smaller tumor size, favorable biologic subtype (ER+) and adequate margins of excision.

The results of our own study are consistent with the above but allow a few additional observations and caveats. First, isolated IBTR at a first event represents a favorable subset of all patients with local recurrence, many of whom present with coincident regional node or distant disease; by excluding the latter, better outcomes are not surprising. Second, lacking fixed criteria, the choice of RCT versus mastectomy represents a composite of surgeon and patient preference, and our surgeons’ thresholds for recommending RCT may have varied; interestingly, aside from obvious selection for RCT on the basis of older patient age and longer DFI, the histopathologic characteristics of the initial cancers were similar between RCT and mastectomy, with BCSS and OS from time of initial surgery virtually identical. Even allowing for our selection biases, these date suggest more similarities than differences in the initial tumor biology, which – along with DFI - remains the main determinant of survival post-IBTR.^5,6,15,16 Third, 90% of isolated IBTR were non-palpable, with the largest proportion detected by mammography, 57–53% being T1, 25–27% being DCIS, 9–16% with no residual cancer at the biopsy site, posing a very limited future risk of systemic relapse; this experience suggests that follow-up by physical exam is insufficient and supports annual breast imaging following BCT, consistent with current guidelines. Fourth, the yield of axillary surgery for IBTR was small, with subclinical nodal metastases found in only 10 of the 192 patients who had nodes removed as part of their procedure; although we²¹ and many others²² have reported the feasibility of re-operative sentinel lymph node biopsy (SLNB) in patients with IBTR, we²³ have more recently questioned its value for patients at sufficient risk of systemic relapse to require systemic adjuvant therapy regardless of node status, and in whom other outcomes (including nodal relapse) seem to be comparable with or without node staging at the time of surgery for IBTR. Fifth, following RCT the largest number of subsequent next events were re-IBTR, occurring in 17/130 (13%) patients, a 5-year rate of recurrence significantly higher than that expected following either a “first time” BCT procedure or a mastectomy for IBTR, 4/192 (2%), but one which does not appear to affect survival. In contrast, most subsequent events following mastectomy were distant. Sixth, our results are limited by a time frame (1999–2019) during which the various techniques of whole-breast RT and – especially in recent years – partial breast RT have evolved, precluding meaningful subset analyses and specific treatment recommendations. Finally, since all prior studies including our own share the limitations of retrospective design and selection bias, we cannot prove equivalence between RCT and mastectomy for IBTR, and a randomized trial – given current low rates of isolated IBTR - seems unlikely. What we and others^3–7,9,19 have clearly shown is that RCT is suitable for a growing proportion of patients with isolated IBTR, with excellent results in properly selected patients.

Conclusion

In a large cohort study comparing RCT with mastectomy for patients with isolated IBTR following BCT, we have observed selection bias on the basis of older age and longer DFI favoring RCT, but comparable BCSS and OS, both from time of initial treatment (at a median follow up of 10.7 years) and from time of IBTR (with a median follow up of 6.7 years). RCT deserves wider consideration in the multidisciplinary management of IBTR, especially in the setting of favorable pathology, older age and longer DFI.