Time Intervals between Double Primary Breast and Ovarian Cancers and Survival Outcomes of Patients with Both Cancers: A SEER Database Analysis.

Background: The time interval rules and survival outcomes of individuals with synchronous and metachronous breast cancer (BC) and ovarian cancer (OC) were examined in this retrospective population-based investigation.
Methods: The National Cancer Institute's Surveillance, Epidemiology, and End Results database was used to create a cohort of people diagnosed with BC and OC between 1973 and 2015. Patients were separated into three groups: those with main BC followed by primary OC (group 1), those with synchronous primary breast and ovarian cancer (group 2), and those with OC prior to BC (group 3). The Kaplan-Meier technique was used to assess overall survival (OS) and cancer-specific survival (CSS).
Results: A total of 4,975 patients were identified: 2,929 patients in group 1, 680 patients in group 2, and 1,366 patients in group 3. The average duration between these tumors was 60 months (range 0-499). Approximately 50% of second primary cancer cases occurred during the first 60 months of the first primary cancer diagnosis, and more than 70% occurred within the first 120 months. The median survival time for 4,975 individuals was 140 months. Group 2 had the smallest median OS (35 months), whereas group 3 had the longest (45 months) (239 months). Conclusions: The majority of second primary cancer cases occurred during the first 120 months following the diagnosis of the first original malignancy. Individuals who had primary OC prior to BC had better prognoses, whereas patients who had synchronous BC and OC had worse prognoses.

1. Introduction

Breast cancer (BC) and ovarian cancer (OC) are the two most common malignancies in women [1-4]. In 2017, about 252,710 new instances of invasive BC and 22,440 new cases of OC were projected among women in the United States [2], and people susceptible to both primary BC and main OC are relatively uncommon clinical entities [5, 6].

Patients who have both BC cancer and primary OC form a subpopulation known as double primary breast and ovarian cancer (DPBOC) [7]. The BRCA1- and BRCA2-linked hereditary breast and ovarian cancer syndromes (HBOC) [8] are among the most well-known and intensively researched hereditary cancer disorders. Previous research has showed that women with BC are more likely to acquire OC and vice versa [9-13]. Bergfeldt et al., for example, found that the risk of OC was significantly elevated in young women with BC who also had a family history of breast or ovarian cancer [9]. Metcalfe et al. discovered that the 10-year actuarial probability of OC following BC for BRCA1 carriers was 12.7 percent and 6.8 percent for BRCA2 carriers [14]. The risk of future ovarian cancer was tenfold enhanced in women with BC with BRCA1 and BRCA2 mutations [15]. Furthermore, the risk of BC was raised in OC patients with BRCA1/BRCA2 mutations, and the incidence of BC varied from 3.9 percent to 10.98 percent in those women [13, 16, 17]. Nonetheless, many people who tested negative for BRCA had harmful mutations in other suppressor genes and oncogenes linked to hereditary breast and/or ovarian malignancies. TP53 in Li-Fraumeni syndrome, PTEN in Cowden syndrome, mismatch repair (MMR) genes in Lynch syndrome, and CDH1 in diffuse gastric cancer syndrome are examples of these genes [18, 19].

It is vital to do research on DPBOC and investigate the clinical features of this subgroup in order to give evidence for accurate cancer prevention. However, population-based clinical features and survival outcome assessments for DPBOC remain uncommon. There is little known regarding the time intervals between initial ovarian cancer and breast cancer diagnosis, as well as the prognosis of people with both cancers. As a result, the current study looked at the clinical features, time intervals between these two main tumors, and survival outcomes of individuals with double primary malignancies in the general population.

2. Patients and Methods

2.1. Data Source

This study used data from the National Cancer Institute's Surveillance, Epidemiology, and End Results (SEER) database, which was released in November 2017. Since 1973, the SEER program has collected data on cancer cases, and in 2015, there were 18 population-based cancer registries in the United States (Detroit, Iowa, Kentucky, Louisiana, Utah, Connecticut, New Jersey, Atlanta, Rural and Greater Georgia, Alaska, California, Hawaii, Los Angeles, New Mexico, San Francisco, San Jose, and Seattle), representing approximately 30 percent of the US general population. To extract data, the SEER program statistical analysis software package (SEER∗Stat version 8.3.5) was employed. To identify all women with BC and OC, the ICD-O-3/WHO 2008 site-specific code, as well as the ICD-O-3 histological code and behavior, were used. SEER∗Stat's “case-listing” option was used to retrieve demographic, clinicopathological, and survival data. Due to the nature of the SEER program as an open-access resource, this study was exempt from institutional review board approval.

2.2. Study Groups

Initially, SEER∗Stat 8.3.5 was used to extract primary BC and primary OC cases between 1973 and 2015, and malignancies that had metastasized to the breast or ovary from another origin were removed from the research. Primary malignant tumors with three or more nodes were also eliminated. Searching for the identical study identification number between the two databases yielded double primary cancer cases [20-23].

Cases with double primary cancer were classified into three categories based on the sequences and time intervals between the two malignancies, as follows: BC before to OC (group 1), synchronous (group 2), and BC after OC (group 3). Two primary tumors with time intervals equal to or less than 6 months between them were classified as synchronous cancer (group 2), whereas cancers with time intervals more than 6 months were classified as metachronous cancers (groups 1 and 3).

2.3. Clinical Information

The SEER database was used to determine the following variables among eligible cases: age at diagnosis, year and month of diagnosis, race/ethnicity, marital status, registration area, patient follow-up status, cancer stage, histological subtype, tumor grade, associated treatment, survival, and cause of death.

Non-Hispanic White, non-Hispanic Black, Hispanic, Asian/Pacific Islander, and other (American Indian/Alaska Native, and unknown race) were the racial/ethnic groups studied. The American Joint Committee on Cancer (AJCC) staging classification scheme was used to determine cancer stage. From 2004 to 2015, OC staging was based on the AJCC 6th edition staging categorization schema [24]. From 1988 to 2015, BC staging was assessed using the AJCC 6th edition tumor node metastasis (TNM) staging classification schema [25]. Tumors were classified as highly differentiated (grade I), moderately differentiated (grade II), poorly differentiated (grade III), undifferentiated (grade IV), or unknown using the International Classification of Diseases for Oncology, 2nd edition (ICD-0-2). The interval period was computed from the month of the first malignant tumor's diagnosis to the month of the second malignant tumor's diagnosis. The time gap between the month of breast or ovarian cancer diagnosis and the month of death from breast or ovarian cancer was designated as cancer-specific survival (CSS).

2.4. Statistical Analysis

Continuous data are shown as the mean, standard deviation (SD), or median (min-max), whereas categorical variables are shown as the number of instances and percentage. Pearson's chi-square or Fisher's exact tests were used to assess nominal variables. For ordinarily distributed continuous data, one-way analysis of variance (ANOVA) was utilized, whereas the Mann–Whitney U test was used for nonnormally distributed continuous variables. The Kaplan-Meier technique was used to create the OS and CSS curves, and the log-rank test was used to compare groups. Statistical significance was defined as P values less than 0.05. The SPSS v.24 statistics program for Windows was used for all statistical analyses (SPSS, Chicago, IL, USA).

3. Results

3.1. Patient Characteristics

In the SEER database, 1,263,266 individuals were classified as having primary BC and 145,344 patients as having primary OC over a 42-year period. From 1973 to 2015, 5,053 of these patients had both main BC and primary OC, accounting for 0.40 percent of patients with primary BC and 3.48 percent of patients with primary OC. After eliminating patients with missing survival and diagnostic time values, this study contained 4,975 individuals. These 4,975 individuals were separated into three groups based on the tumor sequence and time intervals between both cancers: 2,929 in group 1, 680 in group 2, and 1,366 in group 3 (Figure 1). Table 1 lists the characteristics of the patients.

Figure 1

Screening flow chart for patients with primary breast cancer and primary ovarian cancer. Based on the sequence and time intervals between both cancers, double primary cancer cases were divided into the following groups: BC prior to OC (group 1), synchronous (group 2), and BC after OC (group 3).

Table 1

Demographic and clinicopathological characteristics of patients with primary BC and primary OC stratified by tumor sequence and time intervals between both cancers.

Variables	BC to OCGroup 1	SynchronousGroup 2	OC to BCGroup 3	P value	Overall
Total, n (%)	2,929 (58.9%)	680 (13.7%)	1,366 (27.5%)		4,975
Age (median, range) at BC diagnosis	59 (23–93)	64 (28–98)	66 (22–97)	<0.05	62 (22–98)
Age (median, range) at OC diagnosis	67 (25–98)	64 (28–98)	57 (10–90)	<0.05	64 (10–98)
Race				0.008
White	2,372 (81.0%)	526 (77.4%)	1,118 (81.8%)		4,016 (80.7%)
Black	154 (5.3%)	61 (9.0%)	76 (5.9%)		291 (5.8%)
Hispanic	213 (7.3%)	52 (7.6%)	84 (6.1%)		349 (7.0%)
Asian/Pacific islander	176 (6.0%)	35 (5.1%)	84 (6.1%)		295 (5.9%)
Other	14 (0.5%)	6 (0.9%)	4 (0.3%)		24 (0.5%)
OC distribution by year				<0.001
1973–1982	94 (3.2%)	67(9.9%)	245 (17.9%)		406 (8.2%)
1983–1992	349 (11.9%)	84(12.4%)	267 (19.5%)		700 (14.1%)
1993–2002	704 (24.0%)	172 (25.3%)	414 (30.3%)		1,290 (25.9%)
2003–2015	1,782 (60.8%)	357 (52.5%)	440 (32.2%)		2,579 (51.8%)
BC distribution by year				<0.001
1973–1982	439 (15.0%)	67 (9.9%)	57 (4.2%)		563 (11.3%)
1983–1992	613 (20.9%)	85 (12.5%)	164 (12.0%)		862 (17.3%)
1993–2002	1,099 (37.5%)	171 (25.1%)	285 (20.9%)		1,555 (31.3%)
2003–2015	778 (26.6%)	357 (52.5%)	860 (63.0%)		1,995 (40.1%)
OC grade				<0.001
Grade I	99 (3.4%)	46 (6.8%)	156 (11.4%)		301 (6.1%)
Grade II	304 (10.4%)	78 (11.5%)	241 (17.6%)		623 (12.5%)
Grade III	1,013 (34.6%)	201 (29.6%)	432 (31.6%)		1,646 (33.1%)
Grade IV	435 (14.9%)	67 (9.9%)	127 (9.3%)		629 (12.6%)
Unknown	1,078 (36.8%)	288 (42.4%)	410 (30.0%)		1,776 (35.7%)
OC stage (2004–2015)∗	1663	339	382	<0.001	2,384
I	238 (14.3%)	71 (20.9%)	133 (34.8%)		442 (18.5%)
II	144 (8.7%)	26 (7.7%)	49 (12.8%)		219 (9.2%)
III	658 (39.6%)	120 (35.4%)	120 (31.4%)		898 (37.7%)
IV	436 (26.2%)	83 (24.5%)	62 (16.2%)		581 (24.4%)
Unknown	187 (11.2%)	39 (11.5%)	18 (4.7%)		244 (10.2%)
OC pathologic type				<0.001
Serous	1,354 (46.2%)	294 (43.2%)	510 (37.3%)		2,158 (43.4%)
Mucinous	109 (3.7%)	38 (5.6%)	155 (11.3%)		302 (6.1%)
Clear cell	111(3.8%)	28 (4.1%)	85 (6.2%)		224 (4.5%)
Endometrioid	193 (6.6%)	49 (7.2%)	227 (16.6%)		469 (9.4%)
Other	1,162 (39.7%)	271 (39.9%)	389 (28.5%)		1,822 (36.6%)
BC grade				<0.001
Grade I	364 (12.4%)	102 (15.0%)	215 (15.7%)		681 (13.7%)
Grade II	715 (24.4%)	172 (25.3%)	451 (33.0%)		1,338 (26.9%)
Grade III	883 (30.1%)	197 (29.0%)	435 (31.8%)		1,515 (30.5%)
Grade IV	83 (2.8%)	11 (1.6%)	16 (1.2%)		110 (2.2%)
Unknown	884 (30.2%)	198 (29.1%)	249 (18.2%)		1,331 (26.8%)
BC stage (1988–2015)#	2,218	566	1,235	<0.001	4,019
I	1,100 (49.6%)	191 (33.7%)	625 (50.6%)		1,916 (47.7%)
II	710 (32.0%)	157 (27.7%)	343 (27.8%)		1,210 (30.1%)
III	214 (9.6%)	53 (9.4%)	109 (8.8%)		376 (9.4%)
IV	34 (1.5%)	69 (12.2%)	54 (4.4%)		157 (3.9%)
Other	160 (7.2%)	96 (17.0%)	104 (8.4%)		360 (9.0%)
BC pathologic type				0.096
Infiltrating ductal carcinoma	2,091 (71.4%)	458 (67.4%)	976 (71.4%)		3,525 (70.9%)
Other	838 (28.6%)	222 (32.6%)	390 (28.6%)		1,450 (29.1%)
Interval (median, range), in months	75 (7–499)		73 (7–496)		60 (0–499)

Note: ∗Ovarian cancer stage derived from AJCC 6th Stage (2004+); #breast cancer stage adjusted for AJCC 6th Stage (1988+); OC: ovarian cancer; BC: breast cancer.

The most patients were in group 1 (58.9 percent, 2,929/4,975), followed by Group 3 (27.5 percent, 1366/4,975) and group 2 (13.7 percent, 680/4,975). The median age at BC diagnosis was 62 years (range 22–98 years) among the 4,975 double primary cancer patients, while the median age at OC diagnosis was 64 years (range 10–98 years). White women were responsible for the large majority of instances in all three categories.

We identified only the OC case stages from 2004 to 2015 and the BC case stages from 1988 to 2015 since the 6th edition AJCC staging classification schema began in 2004 for OC and 1988 for BC.

Group 3 showed a greater number of early-stage OC patients (stages I and II, 47.6%, 182/382) than groups 1 (23.0%, 382/1,663) and 2 (28.6%, 97/339). Furthermore, group 3 showed a larger proportion of grade I and II OC cases (29.1%, 397/1,366) than groups 1 (13.8%, 403/2,929) and 2 (18.2%, 12 4/680). Group 3 had 37.3 percent fewer OC patients with a serous histological type (510/1,366) than groups 1 (46.2 percent, 1,354/2,929) and 2 (43.2 percent, 294/680). Furthermore, from 1988 to 2015, group 2 had more stage IV BC cases (12.2 percent, 69/566) than groups 1 (1.5 percent, 34/2,218) and 3 (4.4 percent, 54/1,235). The pathological type of most BC patients (70.9 percent, 3,525/4,975) was infiltrating ductal carcinoma.

3.2. Time Intervals

Figure 2 depicts 239 individuals in group 2 who were diagnosed with two primary malignancies in the same month. Up to 50% of patients had shorter than 60-month intervals between malignancies, and more than 70% had less than 120-month intervals. Furthermore, as the interval lengths rose, the number of patients with second cancer reduced. The total median time intervals between both cancers were 60 months (range 0–499 months), with the median intervals for group 1 being 75 months (range 7–499) and group 3 being 73 months (range 7–496).

Figure 2

Time interval distribution between primary breast cancer and primary ovarian cancer.

3.3. Survival Outcome

The median overall survival time for the 4,975 patients was 140 months (95 percent CI: 134.7–145.3). The Kaplan-Meier approach revealed that the median OS for women with synchronous double primary tumors was considerably shorter (P < 0.001) than for those with metachronous double primary tumours (group 1, 134 months, 95 percent CI: 128.4–139.6; group 3, 239 months, 95 percent CI: 218.3–259.7). (Figure 3(a), Table 2).

Figure 3

(a) Overall survival for the three groups after the first malignant tumor. (b) Breast cancer-specific survival. (c) Ovarian cancer-specific survival.

Table 2

Survival outcome.

Variables	BC to OCGroup 1	SynchronousGroup 2	OC to BCGroup 3	P value	Overall
Total, n (%)	2,929 (58.9%)	680 (13.7%)	1,366 (27.5%)		4,975
OS (median, range)	134 (128.4–139. 6)	35 (29.7–40.3)	239 (218.3–259.7)	<0.001	140 (134.7-145.3)
CSS of OC	45 (41.5–48.5)	91 (57.8–124.3)	NC	<0.001	127
Deceased	2,144 (73.2%)	488 (71.8%)	682 (49.9%)	<0.001	3,314 (66.6%)
Died of OC	1,539 (52.5%)	252 (37.1%)	210 (15.4%)	<0.001	2,001 (40.2%)
Died of BC	219 (7.5%)	103 (15.1%)	204 (14.9%)	<0.001	526 (10.6%)
Died of other	386 (13.2%)	133 (19.6%)	268 (19.6%)	<0.001	787 (15.8%)
Alive	785 (26.8%)	192 (28.2%)	684 (50.1%)	<0.001	1,661 (33.4%)

Abbreviations: OS: overall survival; CSS: cancer-specific survival; BC: breast cancer; OC: ovarian cancer; NC: not calculated.

3,314 (66.6 percent, 3314/4975) of the 4,975 women had perished. 526 (10.6 percent, 526/4,975) of the women died of BC, 2001 (40.2 percent, 2001/4,975) died of OC, and 787 (15.8 percent, 787/4,975) died of other reasons. The proportion of fatalities in group 1 was greater (73.2 percent, 2,144/2,929) than in groups 2 (71.8 percent, 488/680) and 3 (49.9 percent, 682/1366) (Table 2). The OC death rate in group 1 (52.5 percent, 1,539/2,929) was clearly higher than that in groups 2 (37.1 percent, 252/680) and 3 (15.4 percent, 210/1,366) (P < 0.001); however, the BC death rate in group 1 (7.5 percent, 219/2,929) was lower than that in groups 2 and 3 (group 2: 15.1%, 103/680; group 3: 14.9%, 204/1,366) (P < 0.001).

Figure 3(b) demonstrates that women in group 1 had greater BC-specific survival than those in groups 2 and 3 (P < 0.001); however, women in group 1 had inferior OC-specific survival than those in groups 1 and 2 (Figure 3(c), P < 0.001).

4. Discussion

Previous research on primary BC and OC focused mostly on epidemiology and cancer susceptibility genes [6–10, 12, 13, 26]. Despite this, nothing is known regarding the time intervals between two initial cancer diagnoses and the prognosis of people with both malignancies. To the best of our knowledge, this is the first population-based study to look into the time intervals and outcomes of individuals with both primary BC and OC.

In the current study, patients were separated into three groups based on tumor sequencing and time intervals between cancer diagnoses: women in group 1 had primary BC followed by primary OC, women in group 2 had synchronous BC and OC, and women in group 3 had OC before BC. Synchronous cancer was defined as the diagnosis of both primary tumors at the same time or within 6 months of each other, whereas metachronous cancer was defined as the diagnosis of both main tumors more than 6 months apart.

However, past research has revealed that there is no agreement on the concept of synchronous cancer. Syncynchronous cancer was characterized by Rose et al. and Lavrador et al. as cancers diagnosed in the same year [27, 28]. Matsuo et al. classified synchronous cancer as a time gap of less than four months between two cancer diagnoses since the great majority of the second cancer was detected during a hysterectomy done within four months of the first cancer diagnosis [20]. In another research, patients with synchronous cancer were classified as those who were diagnosed with a secondary cancer within 6 months of being diagnosed with their initial primary cancer [29]. Based on the assumption that the second main tumor already existing when the first primary tumor was identified, we utilized a 6-month time gap as the cut-off value between both cancer diagnoses.

Several prior research have reported on the time intervals between initial BC diagnosis and OC diagnosis. Bergfeldt et al. discovered that the average period between BC and OC diagnosis was 7 years [9]. Metcalfe et al. [14] found a mean time of 8.1 years (range 0.1–25.5 years) from BC to OC. According to Gangi et al., the median period from EOC diagnosis to BC diagnosis was 50.5 months [17], but McGee et al. observed that the average time from OC diagnosis to BC diagnosis in BRCA mutation carriers was 3.5 years [13]. According to Domchek et al., the median time to BC following OC was 108 months (range, 13–241) [16]. Nonetheless, all of the findings presented above were based on research with limited sample sizes. The median time intervals between both malignancies in the current population-based analysis were 60 months (range 0–499 months), whereas the median time intervals for women in group 1 were 75 months (range 7–499) compared to 73 months (range 7–496) for women in group 3.

Figure 2 depicts the time interval distribution between double primary tumor diagnoses, which varies from the time interval distribution of other double primary cancers [29]. In all, 239 people were diagnosed in the same month with both primary cancers. Approximately 50% of second primary cancer cases occurred during the first 60 months of the first primary cancer diagnosis, and more than 70% occurred within the first 120 months. Shorter time intervals resulted in more women developing second primary cancer. Furthermore, as the time intervals rose, the frequency of women diagnosed with a second malignancy reduced. Our findings were comparable with a recent study, which found that around 52% of second primary cancer cases occurred during the first 5 years following BC diagnosis [11]. Based on the findings, the possibility of a synchronous cancer should be ruled out when diagnosing the initial primary cancer, and it is fair to begin BC surveillance as soon as feasible in patients newly diagnosed with OC. Similarly, starting OC screening as soon as feasible in individuals newly diagnosed with BC seems sensible.

The median OS of 4,975 participants in the current research was 140 months. Furthermore, individuals with synchronous double primary tumors (group 2) had a considerably shorter median survival time than those with metachronous double primary malignancies. Because of the grouping approach, 125 women who died within 6 months of being diagnosed with a malignant tumor were all included in group 2, amounting for 18.4 percent of this group. However, after eliminating 139 women who died within 6 months of the first primary cancer diagnosis or had follow-up durations of less than 7 months, the median OS of group 2 was 51 months, which was still lower than that of groups 1 and 3. One probable explanation for the poor prognosis is that group 2 (12.2 percent, 69/680) comprised more patients with stage IV BC than groups 1 (1.5 percent, 34/2929) and 3 (4.4 percent, 54/1366). Inclusion of these patients also resulted in worse BC-specific survival in group 2 compared to groups 1 and 3 (P < 0.001) (Figure 3(b)). Further research appears to be needed to confirm our findings.

The median OS for group 3 was substantially greater than for groups 1 and 2 (P < 0.001). A recent study found that women who underwent metachronous BC after EOC had a greater overall survival rate than those who just had EOC [17]. Because of the substantial probability of early OC recurrence, advanced-stage patients with short disease-free intervals had little chance of developing BC. As a result, approximately half of the women in group 3 had early-stage cancer; hence, these women may survive for a long time following OC diagnosis before developing breast cancer. Furthermore, more differentiated (11.4 percent, 156/1,366) and less serous histological subtype (37.3 percent, 510/1,366) OC patients may have contributed to group 3's increased survival compared to groups 1 and 2. Based on the foregoing, more than half of the patients in group 3 were still alive in the current research (50.1 percent, 684/1,366).

According to Rose et al., the most virulent of the synchronous tumors defined mortality rates, while the mortality rate of individuals with metachronous tumors was determined by second malignancies after the first neoplasm was cured [28]. Women in group 1 had higher OC death rates and lower BC death rates than those in groups 2 and 3, resulting in inferior OC-specific survival and better BC-specific survival (P = 0.001). The OC, as the second major malignancy, was primarily responsible for the mortality rate in group 1, which was consistent with Rose's assessment of the mortality rate of metachronous tumor. 37.1 percent (252/680) of women in group 2 (synchronous) died from OC. As the most virulent of the synchronous tumors, OC also influenced the death rate in group 2, which agreed with Rose's assessment of synchronous tumors. In group 3, 210 (15.4 percent, 210/1,366) women died of OC, 204 (14.9 percent, 204/1,366) died of BC, 268 (19.6 percent, 268/1,366) died of other causes, and 684 (50.1 percent, 684/1,366) survived. Clearly, neither OC nor BC impacted the death rate of group 3, which contradicted Rose's conclusion about the mortality rate of metachronous cancers.

This study has some positives, but it also has several shortcomings. To begin with, the SEER database does not contain information on BRCA1/2 gene mutations. Furthermore, due to the retrospective character of this study, selection bias may have been introduced.

5. Conclusion

The current study reveals a time interval rule between primary BC and primary OC diagnoses in great detail. Approximately 50% of second primary cancer cases occurred during the first 60 months of the first primary cancer diagnosis, and more than 70% occurred within the first 120 months. Individuals who have primary OC before BC have a better prognosis than patients who have BC followed by OC or simultaneous BC and OC. To avoid and identify the second cancer early, doctors must grasp the time gap rule between twin primary tumors.