Annals of Surgical Oncology

, Volume 14, Issue 5, pp 1638–1643

Results of 23,810 Cases of Ductal Carcinoma-in-situ


  • William E. SumnerIII
    • Division of Surgical OncologyDepartment of Surgery, Sylvester Comprehensive Cancer Center
  • Leonidas G. Koniaris
    • Division of Surgical OncologyDepartment of Surgery, Sylvester Comprehensive Cancer Center
  • Sarah E. Snell
    • Division of Surgical OncologyDepartment of Surgery, Sylvester Comprehensive Cancer Center
  • Seth Spector
    • Division of Surgical OncologyDepartment of Surgery, Sylvester Comprehensive Cancer Center
  • Jodeen Powell
    • Division of Surgical OncologyDepartment of Surgery, Sylvester Comprehensive Cancer Center
  • Eli Avisar
    • Division of Surgical OncologyDepartment of Surgery, Sylvester Comprehensive Cancer Center
  • Frederick Moffat
    • Division of Surgical OncologyDepartment of Surgery, Sylvester Comprehensive Cancer Center
  • Alan S. Livingstone
    • Division of Surgical OncologyDepartment of Surgery, Sylvester Comprehensive Cancer Center
    • Division of Surgical OncologyDepartment of Surgery, Sylvester Comprehensive Cancer Center

DOI: 10.1245/s10434-006-9316-1

Cite this article as:
Sumner, W.E., Koniaris, L.G., Snell, S.E. et al. Ann Surg Oncol (2007) 14: 1638. doi:10.1245/s10434-006-9316-1



Screening mammography has increased the number of patients diagnosed with ductal carcinoma-in-situ (DCIS) in the past 20 years. The Florida Cancer Data System is the largest single source incident cancer registry in the United States. We analyzed this registry to determine the changing incidence and treatment patterns for DCIS.


Patients with DCIS from 1981 to 2001 were identified. Age-adjusted rate, descriptive statistics, and incidence of future DCIS and invasive breast cancer were calculated.


A total of 23,810 DCIS patients were identified. The age-adjusted rate of DCIS has risen from 2.4 to 27.7 per 100,000 women between 1981 and 2001. Median age was 64 years; 85% of patients were white, 6.6% African American, and 7.5% Hispanic. Median tumor size was .9 cm. Forty-seven percent of patients had breast-conserving therapy (BCT). Half of the 53% of patients undergoing mastectomy underwent a modified radical mastectomy. Eight percent received no surgical treatment. Sentinel lymph node biopsy was used in 2.7% of patients who underwent a mastectomy. After BCT, 37.5% received adjuvant radiotherapy, and only 13% were treated with hormonal therapy.


The incidence of DCIS has risen dramatically with the advent of screening mammography. Increasing numbers of these patients are treated with BCT, although a large proportion are still treated with mastectomy, in some cases combined with axillary dissection. Sentinel lymph node biopsy and tamoxifen are important components of therapy, the use of which is slowly increasing in the treatment of DCIS.


DCISBreast cancerIntraductal carcinomaMastectomyBreast conservation

Ductal carcinoma-in-situ (DCIS) is a growing health problem in the United States. Before the advent of screening mammography, the incidence of DCIS was low, and patients presented with DCIS that had become clinically symptomatic. Small series suggested that these patients had a high risk for progression to invasive breast cancer.1 On this evidence, a strategy of aggressive surgical therapy similar to the approach with invasive cancer was adopted. It is unclear whether DCIS detected by screening mammography follows the same natural history.2

Over the past two decades, there has also been a continued shift in the paradigm of treatment for invasive breast cancer favoring breast-conservation therapy, reminiscent of the earlier shift from radical to modified radical mastectomy in the mid-20th century.3 Biopsychosocial issues have also played a greater role in decisions regarding treatment of both invasive cancer and DCIS.4 As additional evidence and new adjunctive therapies become available, these modalities of treatment are being adopted at varying penetrance in different regions of the country, resulting in widely differing practice patterns.58 We analyzed the Florida Cancer Data System to assess the changing incidence of DCIS in Florida and to evaluate evolving practice patterns within the state.


The Florida Cancer Data System is a statewide cancer data registry commissioned in 1978, with data collected and maintained at the University of Miami Sylvester Cancer Center since 1981. We analyzed this registry to evaluate the changing trends in incidence and treatment of women with DCIS within the state of Florida from 1981 through December 2001. Data prospectively collected by the registry include patient demographic information, location and histology of primary lesion, size of primary lesion, method of treatment of the primary site and regional nodal basin, and the use of postoperative radiotherapy, hormonal treatment, or chemotherapy, disease status, cancer status, and date of last contact. These data are captured and reported to a varying degree from individual reporting hospitals within the state according to guidelines set forth for these variables in the Standards for Cancer Registries published by the North American Association of Central Cancer Registries. These data were examined to assess changing incidence and patterns of treatment for DCIS during the study period. Men with breast cancer were excluded from this analysis, as were women with any evidence of invasive breast cancer at the time of treatment for DCIS, such as a focus of microinvasion within the final specimen.

The annual age-adjusted incidence rates for DCIS were calculated and standardized to the U.S. population. Descriptive statistics for the patients were calculated. Rates of therapy adjusted to the total number of DCIS cases per year were calculated, evaluating surgery of the primary site and treatment of the regional nodal basin. In a similar manner, the rate of patients undergoing breast conservation who received postoperative radiotherapy was calculated. Since approval by the U.S. Food and Drug Administration of tamoxifen, the frequency of treatment with hormonal therapy was also evaluated. Categorical variables and nonparametric data were tested for differences by χ2 test. Means of continuous variables were compared with Student’s t-test.


Patient Demographics

There were 23,810 women diagnosed in the state of Florida between 1981 and December of 2001 with DCIS. The median age at diagnosis was 64 years (range, 18–103 years) and did not change across the study period. Most patients were postmenopausal; 80% were older than 50. In 20% of patients, the diagnosis of DCIS was made after age 74. Hispanics made up 7.3% of the patients, and 6.6% were African American. Although 61% percent of patients were married at the time of diagnosis, only 43% (P < .001) of African American patients were married. Thirty-four percent of patients diagnosed with DCIS had smoked.

Trends in Diagnosis and Treatment

DCIS incidence increased from 2.4 cases per 100,000 women in 1981 to 27.7 per 100,000 women by 2001 (Fig. 1). During 1992–2001, there was a 236% increase in the number of women diagnosed with DCIS. The size of the DCIS lesion was sporadically reported. In the 6212 patients for whom this datum was available, the median tumor size was .9 cm. Along with the striking increase in incidence of DCIS during this time period, there has been a marked shift in the treatment paradigm to favor breast-conserving therapy (BCT). Ninety-two percent of patients underwent some form of surgical therapy for DCIS. Overall, 47% of women were treated with BCT. In the first 13 years, an average of 11% of women per year were treated with BCT. After 1993, this number rose precipitously while the percentage of DCIS patients treated with mastectomy decreased (Fig. 2). The same overall increase in use of BCT was mirrored in the Hispanic and African American populations. The number of women who received no surgical treatment for DCIS has remained constant, averaging 7.8% and fluctuating between 6% and 12% across the study period.
FIG. 1.

Incidence of ductal carcinoma-in-situ (DCIS), 1981–2001. Triangles, annual age-adjusted DCIS incidence; diamonds, crude incidence for women ≥50 years of age; circles, crude incidence for women <50 years of age.
FIG. 2.

Percentage of patients with ductal carcinoma-in-situ (DCIS) treated with breast conservation therapy (BCT, diamonds) or mastectomy (total or modified radical, squares) during 1981–2001.

Of the women with DCIS who underwent mastectomy, more recently, fewer axillary dissections have been performed as a component of the patient’s therapy. However, even in 2001, 35% (217 of 627 patients) of the mastectomies performed for DCIS also included an axillary dissection. In these patients, 10 was the median number of lymph nodes collected from the axilla. With the advent of sentinel lymph node biopsy for patients with breast cancer, this modality has also been used in 1% of Florida DCIS patients. Although overall only 2.7% of patients undergoing a total mastectomy for DCIS underwent a sentinel lymph node mapping procedure, there has been a 10-fold increase in the number of patients assessed by sentinel lymph node biopsy (SLNB) in the last 2 years of data, 1.6% in 2000 and 16.7% in 2001. Also, a small percentage (14% in 2001) of those who received a mastectomy and axillary dissection had a SLNB included in the operation. Patients diagnosed with invasive disease were not included in this data set, even if arising in a background of DCIS. As a result of this criterion for patient selection, all axillary nodal basins dissected or sampled during biopsy were found to be free of disease. Overall, 37.6% of patients treated with BCT received postoperative radiotherapy. The number of patients receiving postoperative radiotherapy after BCT rose initially but has reached a plateau during the 1990s (Fig. 3).
FIG. 3.

Percentage of ductal carcinoma-in-situ (DCIS) breast conservation therapy (BCT) patients receiving postoperative radiotherapy (XRT) during 1981–2001.

Overall, 12% of patients had DCIS with comedo type histology. This rate was constant across the study period. Although the presence of comedo histology did not affect the rate of BCT, there was a small but important increase in the likelihood that a patient treated with BCT would be referred for postoperative radiotherapy (37.0% vs. 41.7%, P < .001). Also, among patients undergoing mastectomy for DCIS, those with comedo histology showed a small but statistically significant increase in the number of axillary dissections performed (48.5% vs. 53.3%, P < .033).

We evaluated the number of patients treated with hormonal therapy during the study period. Most of these patients were presumably treated with tamoxifen (which was approved by the U.S. Food and Drug Administration on October 29, 1998) to decrease the rate of recurrence of invasive breast cancer. In 1998, a total of 6.8% of patients with DCIS were treated with hormonal therapy. Since that year, this percentage has increased slowly, to a rate of 14.1% of patients in 2001. Since 1998, patients receiving BCT and radiation were more likely to receive hormonal therapy after surgery than patients undergoing BCT without radiation or mastectomy (28%, 5%, and 7%, respectively; P < .0001). Overall, 7.3% of patients undergoing a mastectomy since 1998 were treated with hormonal therapy, and there was no difference in treatment regardless of whether an axillary dissection was performed.

Survival and Recurrence

With regard to overall survival, although active follow-up is lacking, patients who die within the state are reported to the Florida Cancer Data System by the Florida Department of Vital Statistics. This follow-up is active. Although there is the underlying assumption that patients are not leaving the state and dying, it is possible to estimate an overall survival rate if one assumes that patients not reported as deceased are indeed alive. By using this method, cumulative survival would be 90% at a median follow-up of 101 months. Finally, 648 patients developed a second occurrence of DCIS, and 2517 patients went on to develop secondary breast invasive malignancies.


We used the Florida Cancer Data System to analyze and report our statewide experience with the rapidly increasing pathology of DCIS. This constitutes one of the largest single incident cancer registries in the United States, spanning more than 20 years since inception. Therapy for premalignant and malignant lesions of the breast has undergone marked changes during this time period. This corresponds to the use of increasingly sophisticated technology for the detection and diagnosis of breast lesions, burgeoning experience with early lesions detected by screening, and advances in treatment of breast cancer. Although many questions have been answered regarding the best manner to treat our patients, some remain, and other new questions have been raised.9

Our data concur with other regional reports and with reports from Surveillance, Epidemiology, and End Results program (SEER) data detailing the increase in the incidence of DCIS since the advent in the early 1980s of widespread mammographic screening for breast cancer.10,11 The continued increase in the number of DCIS cases translates into growing costs of care for these patients, both in financial terms and in increased strain on multidisciplinary resources in providing care for a the rapidly increasing numbers of patients.1216 Additional research into the implication of DCIS in development of invasive breast cancer is warranted to justify these expenditures, particularly if we are able to better predict which patients are at increased risk for development of invasive disease.17 The diagnosis of DCIS was made after age 74 in 20% of patients in our study, and the age of diagnosis has not changed during the study period. Although this is older than some other series, this may merely reflect a more elderly population in general within the state of Florida. Depending on the interval between development of a mammographically detectable lesion of DCIS and progression to invasive cancer, screening for and treatment of nonpalpable DCIS beyond a given age may not improve disease-specific or disease-free survival in patients with DCIS.

Most patients diagnosed with DCIS did undergo some form of surgical therapy. The type of therapy, however, varied from BCT with no assessment of regional nodes to patients treated with modified radial mastectomy. Although there was a marked shift over time favoring BCT, there remained a disturbing number of patients potentially overtreated with modified radical mastectomy. This pattern persisted into 2001, the last year of the study period. Although DCIS without a focus of microinvasion is not a strict indication for SLNB, SLNB is gaining a more defined role1820 in patients with DCIS who will undergo a total mastectomy as their initial surgical therapy. An increasing number of total mastectomy patients had their regional nodal basin assessed by SLNB. This procedure was used in some patients who underwent modified radical mastectomy as well, which may represent surgeons becoming credentialed by performing a SLNB followed by modified radical mastectomy.

The incidence of DCIS with comedo necrosis, which is thought to carry an increased risk for development of invasive cancer, has remained constant at 12% despite rising overall numbers of DCIS patients. Statistically, this would result in a higher overall number of patients diagnosed with DCIS and a favorable histology. One would expect that with increased detection of breast cancer through screening, one would see a similar increase in the number of high-grade DCIS lesions. Our data suggest that high-grade DCIS does not represent a progression from low-grade DCIS, but rather a different type of neoplasm occurring at a constant rate. This same prevalence of DCIS with comedo features was observed in an analysis of DCIS data published from SEER.21 This suggests that patients with low-grade lesions such as atypical ductal hyperplasia, lobular carcinoma-in-situ, and low-grade DCIS are at increased risk of developing high-grade or invasive malignancies. Further studies using genomics may help identify distinctly different tumor signatures in low- and high-grade DCIS.

Evidence supports the use of postoperative radiotherapy as a component of BCT to decrease the chance of recurrent DCIS and development of breast cancer in the future,2228 particularly in patients with unfavorable tumor size, grade, and histology. Although referral for postoperative radiotherapy increased with the increase in BCT, it now seems to have reached a plateau. The reason for this plateau is unclear, and it may still change with increased evidence for radiation as a means of better local control. Patients whose histology included comedo features were more likely to receive a more aggressive surgical approach and to receive postoperative radiation. This may be related to the practice of performing triage on DCIS cases on the basis of classifications such as the Van Nuys score,2931 then using this to determine treatment strategies. BCT patients who were treated with postoperative radiation were also more likely to be treated with hormonal therapy, suggesting that surgeons who referred patients for radiotherapy may be more likely to treat patients with hormonal therapy. Continued evaluation will be necessary to assess the impact of recent evidence supporting the use of both postoperative radiotherapy and hormonal therapy as adjuncts to decrease later development of invasive breast cancer.

In conclusion, concurrent with the rapid increase in incidence of DCIS in the United States, the paradigm for treatment of this pathology has shifted dramatically. There remains a broad variability in choice of treatment even within the state of Florida. Some patients are still treated aggressively with surgery, although the predominant surgical therapy for local control is now BCT. Sentinel node biopsy is increasing in the treatment of DCIS. With increased BCT, more patients are receiving postoperative radiation, and patients are increasingly treated with adjuvant hormonal therapy. New evidence from randomized trials continues to accrue and will likely further influence the changing practice patterns used in the management of patients with DCIS.

Copyright information

© Society of Surgical Oncology 2007