Prognosis for patients with metastatic breast cancer who achieve ‘no-evidence-of-disease’ status after systemic or local therapy

Abstract

Purpose

We sought to determine outcomes for patients with metastatic breast cancer (MBC) with no evidence of disease (NED) after treatment and to identify factors predictive of outcome once NED status was attained.

Methods

We reviewed 570 patients with MBC consecutively treated between January 2003 through December 2005. Ninety patients (16%) attained NED, as defined by a complete metabolic response by positron emission tomography or sclerotic healing of bone metastases on computed tomography or magnetic resonance imaging. Median follow-up for patients attaining NED was 100 months (range, 14–134 months).

Results

The 3- and 5-year overall survival (OS) rates for the entire group were 44% and 24%, respectively, compared to 96% and 78% for those attaining NED. Using a landmark analysis, NED status was significantly associated with survival at 2 years (P<0.001, HR 0.23, 95% CI 0.16-0.34) and 3 years (P<0.001, HR 0.20, 95% CI 0.14-0.30). From the time of NED, the median survival was 102 months (range, 14-134) with a 5-year OS and progression free survival (PFS) of 77% and 40%, respectively. On multivariate analysis, HER2+ was significantly associated with OS (vs. ER+, P=0.02, HR 0.44, 95% CI 0.21-0.90), and trastuzumab use was significantly associated with PFS (P=0.007, HR 0.48, 95% CI 0.28-0.82). Thirty-one patients (34%) with NED remained in remission at last follow-up.

Conclusions

MBC patients who attain NED status have a significantly prolonged survival with a durable response to therapy. Ultimately, this study provides essential outcomes data for clinicians and patients living with MBC.

Introduction

Among patients with metastatic breast cancer (MBC) there is a unique subset with disease that shows a sustainable, complete response to therapies that are offered.^1-5 Prior studies have reported that 5%-10% of patients with MBC survive for more than 5 years, and 2%-5% survive for more than 10 years, with many achieving complete responses to therapy despite having had limited options for systemic therapy when those studies were done.^6,7 This albeit uncommon but distinctive subset challenges the belief that MBC is universally fatal.^8,9

Counseling NED patients with MBC about outcomes and expectations is challenging because the time to recurrence after achieving the status of ‘no evidence of disease’ (NED), as well as predictors of shorter NED intervals and the possibility of achieving second NEDs, have not been well studied. From the patient's perspective, this uncertainty is psychologically distressing.^10-12 Studies suggest that lack of insight into disease status and prognosis can lead to poorer outcomes among patients with advanced or incurable cancer, with psychological consequences including increased mistrust and feelings of abandonment.^12,13 Further, the “cancer-free” implication of an NED status is clearly a strongly desired outcome for any patient with MBC, and validation of the frequency with which this status occurs is influential.

The few studies that report outcomes for patients with MBC who achieve a complete response are retrospective reviews of patients treated with older chemotherapeutic regimens and often include only patients with locoregional recurrences, which are biologically different from M1 disease.^7,14,15 Therefore, in an era where long-term survival may be more achievable for patients with MBC because of the use of modern systemic therapies and more aggressive local treatments, we sought to identify and characterize a subset of patients with MBC who attain NED status, to assess their outcomes compared with patients who do not attain NED status, and to identify potential predictors of more durable complete responses.

Methods

Patients

We identified 570 patients with histologically confirmed MBC consecutively treated at The University of Texas MD Anderson Cancer Center who presented between January 2003 and December 2005. Patients had either de novo MBC or initial stage III disease with distant recurrences that appeared during that interval. All staging studies were prescribed per National Comprehensive Cancer Network (NCCN) recommendations.

Medical records were reviewed in detail after obtaining approval from our institutional review board. The site of first distant metastasis was categorized by organ involvement (e.g., lung vs. liver vs. bone); patients were coded as having mixed disease if more than one organ was involved.

NED Subset

Patients were identified who had an interval of no clinical or radiographic evidence of disease. The frequency and modality of radiographic imaging was at the discretion of the treating physician. NED was defined as a complete response by RECIST criteria on computed tomography (CT) and/or a complete metabolic response on positron emission tomography/CT (PET/CT) as interpreted as NED by the reading radiologist. In the setting of bone metastasis, sclerotic bone lesions on CT were considered to have had a complete response if confirmed to be metabolically inactive on PET/CT or without radiotracer uptake on bone scan. Patients for whom review and interpretation of images by MD Anderson radiologists were not available were not classified as having NED.

Follow-up and Statistical Analysis

The median follow-up intervals from the date of distant metastatic disease were 27 months for all patients (range, 0-134 months) and 100 months (range, 14-134 months) for patients with NED.

Descriptive statistics were used to evaluate baseline characteristics, and categorical data were analyzed by using Fisher's exact test and chi-squared analyses. Survival times were calculated from the date of distant metastatic disease. Additional survival times were calculated for the NED subset from the date of NED to the first occurrence of the considered event. NED intervals were calculated from the radiographic date of NED to the time of progression. The Kaplan-Meier method was used to estimate overall survival (OS) for the entire cohort and progression-free survival (PFS) for the NED subset (last reviewed July 9, 2014). Time to NED was analyzed by using the methods of Fine and Gray;¹⁶ for patients who did not achieve NED, death was treated as a competing risk. Log-rank tests were applied to assess for equality across groups for all analyses. A two-sided significance level of 5% was used for analysis.

The Cox proportional hazard model was used for univariate and multivariate analysis to assess the potential influence of patient, tumor, and other factors on the endpoints. Multivariate assessment was performed in a stepwise fashion, with only factors having a P value of ≤0.05 included in the model. Estimated hazard ratios (HRs) were reported when significant (P≤0.05). IBM SPSS Statistics 22 and SAS/STAT were used for data analysis.

Results

Patient Characteristics: All Patients

Patient and tumor characteristics are summarized in Table 1. Ninety patients with MBC (16%) attained NED status. The median time to NED among those who attained it was 11 months (range, 1-43 months). Patients with NED were more likely to have normal body mass index (BMI; i.e., 18.5-24.9) (P<0.001), have ER+ or HER2+ tumors (P<0.001), present with de novo stage IV disease (P=0.04), and have single sites of metastasis (P<0.001) compared with the non-NED group.

Table 1. Patient and Disease Characteristics.

Variable	Value or No. of Patients
	All Patients (n=570)	NED (n=90)	Not NED (N=480)	P Value
Follow-up time, months
Median	27	100	23	0.05
Range	0-134	14-134	0-131
Age, years
Median	50	49	50	0.08
Range	23-91	25-82	23-91
Age Categories, years
<35	49 (9)	10 (11)	39 (8)	0.41
35-50	244 (43)	40 (44)	204 (43)
51-70	229 (40)	36 (40)	193 (40)
>70	48 (8)	4 (4)	44 (9)
Sex
Female	568 (100)	90 (100)	478 (99)	0.54
Male	2 (0)	0 (0)	2 (1)
Race
Caucasian	367 (64)	68 (76)	299 (62)	0.14
Black	115 (20)	11 (12)	104 (22)
Hispanic	64 (11)	8 (9)	56 (12)
Asian	19 (3)	3 (3)	16 (3)
Other	5 (1)	0 (0)	5 (1)
Menopausal Status
Pre	255 (45)	50 (56)	205 (43)	0.08
Post	309 (54)	40 (44)	269 (56)
Unknown	6 (1)	0 (0)	6 (1)
Body Mass Index
Median	28	28	29	0.54
Range	16-107	19-47	16-107
Body Mass Index Categories
Underweight	8 (1)	0 (0)	8 (2)	<0.001
Normal	151 (26)	40 (44)	111 (23)
Overweight	169 (30)	25 (28)	144 (30)
Obese	239 (40)	25 (28)	205 (43)
Unknown	12 (2)	0 (0)	12 (2)
Disease Stage at Diagnosis
IIIA	74 (13)	11 (11)	63 (13)	0.17
IIIB	124 (22)	16 (18)	108 (23)
IIIC	48 (8)	3 (3)	45 (9)
IV	324 (57)	60 (67)	264 (55)
Stage IV
At diagnosis	324 (57)	60 (67)	264 (55)	0.04
At recurrence	246 (43)	30 (33)	216 (45)
Inflammatory at Diagnosis
Yes	103 (18)	16 (18)	88 (18)	0.94
No	467 (82)	74 (82)	393 (82)
Histology
Invasive Ductal	441 (77)	71 (79)	271 (77)	0.94
Invasive Lobular	44 (8)	6 (7)	38 (8)
Mixed	34 (6)	6 (7)	28 (6)
Sarcoma	10 (2)	2 (2)	8 (2)
Other	11 (2)	2 (2)	9 (2)
NOS	30 (5)	3 (3)	27 (6)
Receptor Status
ER+/Her2-	250 (44)	44 (49)	206 (43)	<0.001
Her2+	161 (28)	37 (41)	124 (26)
TNBC	122 (21)	7 (8)	115 (24)
Unknown	37 (6)	2 (2)	35 (7)
Site of First Distant Metastasis
Bone	144 (25)	28 (31)	116 (24)	<0.001
Lung	39 (7)	9 (10)	30 (6)
Liver	64 (11)	19 (21)	45 (9)
Brain	17 (3)	0 (0)	17 (4)
LN/Contralateral Breast	59 (10)	17 (19)	42 (9)
Other (skin, pleura, etc)	15 (3)	0 (0)	15 (3)
Mixed (2+)	232 (41)	17 (19)	215 (45)
PET scan Used for Response
Yes	91 (16)	32 (36)	59 (12)	<0.001
No	479 (84)	58 (64)	421 (88)
Obtained NED Status
Yes	90 (16)
No	480 (84)

Abbreviations: NED, no evidence of disease; NOS, not otherwise specified; ER+, estrogen receptor positive; Her2, human epidermal growth factor receptor 2; TNBC, triple negative breast cancer; LN, lymph node.

Patients attaining NED were also more likely to have received PET/CT imaging than patients not attaining NED (P<0.001). The majority of the PET/CT scans in patients attaining NED (78%) were used to confirm a metabolic complete response in lymph nodes, in sclerotic bone lesions or in residual post-therapy tissue on CT.

Survival and Predictors of Survival: All Patients

For the entire group of 570 patients, the 1-, 3-, and 5-year actuarial OS rates were 77%, 44%, and 24%, with a median survival time of 31 months (Figure 1).

Figure 1.

Kaplan-Meier curves depicting survival from the time of distant metastasis: A) All patients, B) survival of patients attainting NED compared to those not attaining NED, and C) survival based on presenting stage (P<0.001).

Several patient and tumor characteristics influenced survival. Patients with a normal BMI had longer median survival times than did patients with an overweight (25-29.9) or obese (≥30) BMI (40 vs. 32 vs. 24 months, P<0.001). Patients who had de novo MBC had longer median survival times than patients who initially had stage III disease but later had distant recurrence (38 vs. 19 months, P <0.001). Furthermore, for patients who initially presented with stage III disease, the stage subtype predicted median survival upon metastatic recurrence (stage IIIA 23 months vs. stage IIIC 9 months, P<0.001) (Figure 1). On multivariate analysis, several significant factors associated with better OS included: presenting with de novo stage IV disease (P<0.001, HR 0.45) and having local therapy of the primary (P=0.04, HR 0.76) followed by adjuvant radiation therapy (P<0.001, HR 0.55) (Table 2). A second multivariate model was constructed with NED as a time dependent covariate, in which NED status was non-significant.

Table 2. Multivariate Analysis of Factors Associated with Overall Survival.

Variable	Multivariate for OS
	P value	HR	95% CI
Race
Non-black	Reference
Black	<0.001	1.66	1.30-2.12
Body Mass Index
Continuous	0.008	1.017	1.004-1.03
Stage at presentation
III	Reference
IV	<0.001	0.45	0.33-0.62
Receptor Status
ER+	Reference
HER2+	0.60	0.94	0.73-1.19
TNBC	<0.001	3.05	2.30-4.05
Site of DM
Mixed	Reference
Bone	<0.001	0.55	0.42-0.70
Lung	<0.001	0.45	0.30-0.67
Liver	0.03	0.70	0.50-0.97
Brain/Other	0.005	0.65	0.49-0.88
Adjuvant Hormones for Primary
No	Reference
Yes	0.008	1.61	1.14-2.30
Local Therapy of Primary
No	Reference
Yes	0.04	0.76	0.59-0.98
Adjuvant RT to Primary
No	Reference
Yes	<0.001	0.55	0.40-0.79

Abbreviations: OS, overall survival; ER+, estrogen receptor positive; HER2, human epidermal growth factor receptor 2; TNBC, triple negative breast cancer; DM, distant metastasis; RT, radiation therapy.

Using a multivariate landmark analysis, attaining NED status was significantly associated with survival at 2 years (P<0.001, HR 0.23, 95% CI 0.16-0.34) and 3 years (P<0.001, HR 0.20, 95% CI 0.14-0.30).

Factors Associated with Attaining NED

To identify factors that affect the likelihood of attaining NED status, we used cumulative incidence plots to analyze time to NED, with an HR <1.0 representing variables that are less likely to be linked with NED (Table 3). After adjusting for confounders, the variables that remained associated with a decreased likelihood of NED on multivariate analysis were overweight (P=0.01, HR 0.52) and obese BMI (P=0.001, HR 0.43) and TNBC (P=0.005, HR 0.32), whereas presenting with de novo MBC (P<0.001, HR 2.49), having a single metastatic site versus multiple (all ≤0.01, HR ≥2.0), and having had local treatment of the primary tumor (P<0.001, HR 4.1) were associated with increased likelihood of NED (Table 3).

Table 3. Univariate Analysis and Multivariate Analysis of Factors Associated with Time to No Evidence of Disease using the methods of Fine and Gray.

Variable	Univariate Analysis			Multivariate Analysis
	P value	HR*	95% CI	P value	HR*	95% CI
Age at Metastatic Diagnosis
Continuous	0.02	0.98	0.97-0.99	-	-	-
Race
Non-black	Reference
Black	0.04	0.52	0.28-0.97	-	-	-
Body Mass Index
Normal	Reference
Overweight	0.02	0.55	0.33-0.90	0.01	0.52	0.31-0.86
Obese	<0.001	0.39	0.24-0.65	0.001	0.43	0.26-0.72
Menopausal Status
Premenopausal	Reference
Postmenopausal	0.03	0.64	0.42-0.96	-	-	-
Disease Stage at Presentation
III	Reference
IV	0.04	1.57	1.01-2.43	<0.001	2.49	1.53-4.05
Receptor Status
ER+	Reference
HER2+	0.17	1.35	0.88-2.10	-	-	-
TNBC	0.004	0.31	0.14-0.69	0.005	0.32	0.14-0.71
Adjuvant Chemo for Primary
No	Reference
Yes	0.06	0.56	0.30-1.02	-	-	-
Adjuvant Hormones for Primary
No	Reference
Yes	0.12	0.62	0.33-1.14	-	-	-
Local Therapy of Primary
No	Reference
Yes	<0.001	3.12	1.78-5.50	<0.001	4.08	2.23-7.49
Site of Distant Metastasis
Mixed	Reference
Liver	<0.001	4.63	2.42-8.87	<0.001	3.10	1.81-6.37
Lung	0.003	3.28	1.49-4.22	0.07	2.07	0.94-4.59
Bone	<0.001	2.96	1.62-5.35	0.01	2.17	1.17-4.05
Brain/Other	0.005	2.58	1.33-4.98	0.03	2.21	1.11-4.40

Abbreviations: ER+, estrogen receptor positive; HER2, human epidermal growth factor receptor 2; TNBC, triple negative breast cancer.

^*Using cumulative incidence plots, HR <1.0 represent variables that are less likely to be linked with NED.

Relapse and Treatment Characteristics: NED Subgroup

Among 90 patients who attained NED status, 31 (34% of the NED cohort; 5% of the entire population) were NED at last follow-up. Twenty-seven patients (30%) remained NED after the first remission. Of the 63 patients who had recurrent disease, 20 attained a second NED, and 16 of those patients had a second relapse. Within the subset of patients with NED, few differences were found between patients who had recurrent disease after NED and those who did not (Table 4), particularly in number of sites of involvement (P=0.87) or number of distant metastases (P=0.22).

Table 4. Characteristics of Patients Attaining NED Who Did or Did Not Have Recurrence.

Variable	Value or No. of Patients (%)			P Value
	NED (n=90)	No Recurrence (n=27)	Recurrence (n=63)
Follow-up time, months
Median	100	105	97	0.46
Range	14-145	14-129	14-134
Time to NED Status, months
Median	11	13	11	0.53
Range	1-43	3-34	1-43
Duration of NED Status, months
Median	40	93	26	0.39
Range	0-115	0-115	2-89
Stage IV
At diagnosis	60 (67)	19 (70)	41 (65)	0.63
At recurrence	30 (33)	8 (30)	22 (35)
Inflammatory at Diagnosis
Yes	16 (18)	1 (4)	15 (24)	0.02
No	74 (82)	26 (96)	48 (76)
Receptor Status
ER+/HER2-	44 (49)	10 (37)	34 (54)	0.23
HER2+	37 (41)	15 (56)	22 (35)
TNBC	7 (8)	1 (4)	6 (10)
Unknown	2 (2)	1 (4)	1 (2)
Number of Sites Involved
1	73 (81)	23 (85)	50 (79)	0.87
2	12 (13)	3 (11)	9 (14)
3	4 (4)	1 (4)	3 (5)
4+	1 (1)	0 (0)	1 (2)
Number of Mets
1	26 (29)	7 (26)	19 (30)	0.22
2-3	14 (16)	7 (26)	7 (11)
4+	30 (33)	6 (22)	24 (38)
Lymph Nodes	20 (22)	7 (26)	13 (21)
Recurrence Site after NED Status
Same as previous DM	14 (16)
New Site	49 (76)
Treatment of DM to Achieve NED Status
Local Treatment
Yes	35 (39)	13 (48)	22 (35)	0.24
No	55 (61)	14 (52)	41 (65)
Systemic (chemo/Ab)
Yes	67 (74)	23 (85)	44 (70)	0.13
No	23 (26)	4 (15)	19 (30)
Hormone Therapy
Yes	55 (61)	14 (52)	41 (65)	0.24
No	35 (39)	13 (48)	22 (35)
Trastuzumab Use
Yes	37 (41)	15 (56)	22 (35)	0.07
No	53 (59)	12 (44)	41 (65)

Abbreviations: NED, no evidence of disease; ER+, estrogen receptor positive; HER2+, human epidermal growth factor receptor 2; TNBC, triple-negative breast cancer; DM, distant metastasis

Of those attaining a first NED status, 67 patients (74%) received chemotherapy as part of the treatment of MBC to obtain NED status including: taxanes (n=45, 67%), platinum-based agents (n=38, 58%), and doxorubicin (n=24, 36%). Fifty one of those patients (76%) received combination chemotherapy, while 16 received single agent therapy typically with microtubule inhibitors (n=11) or capecitabine (n=3). Fifty-five patients (61%) received hormone therapy (HT), and all 37 of the patients with HER2+ disease (41%) received trastuzumab (Table 4). Thirty-five patients (39%) had local treatment of distant disease, with 16 (46%) receiving radiation therapy (median dose 50 Gy; range, 30-66 Gy) and 19 receiving surgery. The most common distant sites treated with surgery were liver (n=8), bone (n=5), and lung (n=4).

Of the patients attaining a second NED status, 5 were treated with local therapy alone (RT = 4, surgery =1), 4 received combination local and systemic treatment, and 11 were treated with systemic therapy (chemotherapy = 4, HT = 3, or a combination = 4).

Maintenance Therapy following NED Status

After attaining NED, nearly all patients (n=84, 93%) continued maintenance therapy. Of the 65 patients with ER+ disease, 55 (85%) received maintenance HT (while 6 received trastuzumab and 3 received chemotherapy). Thirty seven of those patients (67%) either discontinued the HT at the time of recurrence (n=16) or switched to a different HT (n=21). Nine patients continued maintenance HT indefinitely despite remaining NED, and 9 patients discontinued HT after a median of 60 months (range, 12-60) resulting in 4 relapses.

Of the 37 HER2+ patients, 34 received maintenance (post-NED) trastuzumab (the 3 others received HT). Trastuzumab was commonly continued until cardiac toxicity was documented. The median duration of use was 61 months (range, 8-132) with 7 patients discontinuing at the time of recurrence and 12 continuing beyond first relapse. Six patients remained on trastuzumab (median use, 120 months) with continued NED status, and 9 patients discontinued trastuzumab use resulting in 3 relapses.

A total of 15 patients also received maintenance chemotherapy, most commonly capecitabine (n=8) or paclitaxel (n=4), often in combination with trastuzumab or HT. Three of the patients receiving maintenance chemotherapy had TNBC, while the other 4 patients with TNBC did not receive any maintenance.

Outcomes and Predictors: NED Subgroup

The 1-, 3-, and 5-year OS rates from the time of metastasis for the NED subgroup were 100%, 96%, and 78%, respectively (Figure 1). Notably, patients who attained NED had longer median survival times (112 vs. 24 months among those not attaining NED, P<0.001), though direct comparisons are biased because of guaranteed time.

From the time of NED, the median survival was 102 months (range, 14-134). For patients attaining NED, the probability of survival was as follows: 98% at 1 year (95% CI 91%-99%), 89% at 3 years (95% CI 80%-94%), 77% at 5 years (95% CI 67%-85%), and 39% at 10 years (95% CI 24%-54%).

Having either HER2+ (median not reached) or ER+ disease (90 vs. 23 months for TNBC, P<0.001) was significantly associated with longer survival from the time of NED. On multivariate analysis, age (continuous) (P=0.03, HR 0.97, 95% CI 0.94-0.99), HER2+ disease (vs. ER+, P=0.02, HR 0.44, 95% CI 0.21-0.90), and TNBC (vs. ER+, P<0.001, HR 7.35, 95% CI 2.78-19.4) remained significantly associated with OS from the time of NED.

The PFS rates (after NED) at 1, 3, and 5 years were 87%, 54%, and 40% (Figure 2). The median PFS time after the initial NED status was 41 months (range, 0-115 months), and the median interval from recurrence to a second NED status was 12 months (range, 4-31 months). The only patient-, tumor-, or treatment-related characteristics that were associated with PFS after NED were receptor status (HER2+ 58 months vs. ER+ 35 months vs. TNBC 13 months, P=0.02), inflammatory breast cancer at presentation (28 vs. 45 months, P=0.02), and trastuzumab use (73 vs. 31 months, P=0.01) (Figure 2). Although not statistically significant, directed local therapy for distant metastases resulted in a numerically longer median PFS time (58 vs 31 months, P=0.17). On multivariate analysis, trastuzumab use (P=0.007, HR 0.48, 95% CI 0.28-0.82) and inflammatory breast cancer (P=0.005, HR 2.34, 95% CI 1.29-4.25) remained significant predictors of PFS.

Figure 2.

Kaplan-Meier curves of progression free survival (PFS) from the time of NED: A) All patients and B) PFS stratified by trastuzumab use (P=0.01).

Discussion

In this large retrospective study, we investigated characteristics of patients with MBC in the modern era of targeted systemic treatment. Sixteen percent of our group of 570 patients attained NED. Having a normal BMI, ER+ tumor, de novo MBC, single site of metastatic disease, and local treatment of the primary tumor were all associated with attaining NED Once NED, most patients remained on maintenance therapy, and importantly, patients who attained NED had a 77% probability of surviving 5 years with many having durable remissions that were associated with trastuzumab use. Ultimately, this study provides valuable outcome and prognostic information for the practicing clinician and the patients living with MBC.

The proportion of patients with MBC who attained NED after having been treated from 2003 through 2005, 16%, was similar to that in older studies, including an MD Anderson series of almost 1600 patients in which reported complete response rates were 11%-17%, although arguable advances in imaging may impact this direct comparison.^7,15,17 The similarity in these complete response rates between different treatment eras suggests that perhaps some subset of MBC is biologically more susceptible to treatment despite newer therapies. However, Greenberg and colleagues noted that only 3% of their patients remained NED at 5 years compared with 12% (66 patients) in our cohort. Therefore, even though initially obtaining a complete response may not depend on the use of newer agents, the higher rate of NED at 5 years in this modern series suggests that newer agents have an important role in maintaining the durability of the response.

Given the uncertain outcome among patients who attain NED, this series provides critical prognostic information for patients and their caregivers. We observed a 77% probability of survival at 5 years from the time of NED, which is notably high given the pessimistic view often associated with metastatic disease. This may be in part due to trastuzumab, an agent with demonstrated survival benefit for patients with HER2-positive MBC.¹⁸ For patients who attain NED, trastuzumab may be important in suppressing micrometastases, as suggested by the association between HER2-positive disease and OS and the association between its use and PFS. Also, a higher proportion of patients who remained NED received this agent.

Interestingly, NED status was an independent predictor of survival at 2 and 3 years. This series also suggests that with increasing targets and use of targeted therapy, there may be a higher proportion of MBC patients maintaining NED. Therefore, research will become increasingly necessary to better define this subset. For instance, the biologic behavior of MBC is diverse and undoubtedly influences the ability to attain NED status. Similar to the findings of Dawood and colleagues, among the entire cohort we observed superior survival in patients with de novo MBC (which also was associated with achieving NED status) compared to patients with stage III disease that recurred distantly.¹⁹ Possibly distantly recurrent MBC is chemoresistant because of exposure to previous adjuvant treatment, in contrast to chemotherapy-naïve de novo MBC. Although plausible, this theory does not adequately explain our observation that among patients who presented originally with stage III disease, the initial presenting substage (e.g., IIIA vs. IIIB vs. IIIC) predicted the ultimate outcome at the time of metastasis. Regardless of the cause of this observation, the biologic corollary may be the subject of future preclinical and clinical studies and may imply that de novo MBC should be considered as a unique entity to recurrent stage IV disease.

Similar questions pertain to our observation that one site of disease was associated with higher likelihood of NED and improved survival, regardless of the number of tumor deposits within the visceral organ or bone. A recent study of more than 700 patients noted the same association between survival and the number of organs containing metastases; in that study, for each additional site of disease, the risk of death increased by 18%.²⁰ While this approach to categorizing tumor burden may be simplistic given the heterogeneity among MBC patients, there is no standard classification system. Furthermore, although the prognostic benefit of single-organ involvement is not included in the classic definition of oligometastatic disease, perhaps the number of organs involved is a surrogate of tumor plasticity. Because acquiring the capacity for organ-specific colonization has a temporal component, multivisceral involvement may indicate that micrometastatic disease has been present for longer periods, leading to poorer outcome. Also, genetic diversity is required to colonize dissimilar organ microenvironments; this diversification may promote the emergence of a clonogenic population resistant to therapy, leading to poorer outcomes when multiple sites are involved.²¹ Despite some uncertainty as to the explanation, our observation that single-organ involvement portends better outcomes suggests that clinicians may need to expand their definition of oligometastatic disease for MBC.

We attempted to thoroughly analyze a large cohort of consecutively treated patients with MBC in the modern era, with access to targeted therapy. One important question is whether our results are generalizable. Given that our clinicians use similar treatment schemes as recommended by the NCCN and the 5-year survival rate observed among our MBC patients is similar to that reported using SEER data (24% vs. 25%), there is evidence supporting our experience does correspond to that of the general MBC population.

Despite the strengths of our study, several limitations must be considered when interpreting the results. First, as is true for any retrospective study, inherent bias was present. Second, several studies have implicated performance status as being a major prognostic factor in MBC; we did not have access to these data to incorporate into our survival model. Third, we also did not have information on the therapeutic approach used to treat metastatic disease among patients who did not attain NED status, which limits our ability to comment on treatment approaches that may result in a higher incidence of NED status or how many lines of therapy it took to attain NED. Despite these limitations, our results are highly provocative; future prospective studies, and large database studies, should be used to further validate and explore our results.

In conclusion, the results of our study indicate that patients with MBC who attain an NED status have a significantly prolonged survival with a durable response to therapy. Ultimately, this study provides findings to encourage further research on this subset of patients with MBC, and it provides a backbone of outcome data for clinicians to use when counseling patients who attain complete responses to treatment regarding potential outcomes.