Julliette M. Buckley

The Safety of Multiple Re-excisions after Lumpectomy for Breast Cancer

BackgroundBreast cancer patients may undergo multiple re-excisions after lumpectomy in an attempt to obtain clear margins and avoid mastectomy. We sought to determine the overall local recurrence rate and surgical outcome of patients undergoing two or more re-excisions and to identify predictors of success in attaining clear margins.MethodsRetrospective review of breast cancer patients who underwent lumpectomy for invasive cancer or ductal carcinoma in situ (DCIS) from 1997 to 2007. Patients who underwent two or more re-excisions were identified and analyzed.ResultsWe identified 3,737 patients who underwent lumpectomy over this 10-year period. 875 (23.4%) had close or positive margins requiring a second procedure; 797 (91.1%) had a re-excision; and 78 (8.9%) went directly to mastectomy. Seventy patients underwent multiple re-excisions; 66 patients had 2 re-excisions, 3 patients had 3 re-excisions, and 1 patient had 4 re-excisions. 70% (49/70) of multiple re-excision patients achieved clear margins (26 DCIS, 35 T1, 8 T2, and 1 T3 tumors). All 49 patients who successfully treated with multiple re-excisions received radiation. At a median follow-up of 64 months, 1 of 49 (2.0%) patients had an in-breast recurrence, and 1 of 49 (2.0%) patients had a distant recurrence. Statistically significant risk factors for persistently involved margins after two re-excisions included multifocality and positive lymph node status.ConclusionMultiple re-excisions to obtain clear margins are a safe alternative to mastectomy for women with invasive cancer or DCIS. There is an acceptably low risk of local and systemic failure when negative margins are ultimately achieved.

The feasibility of using natural language processing to extract clinical information from breast pathology reports.

Objective: The opportunity to integrate clinical decision support systems into clinical practice is limited due to the lack of structured, machine readable data in the current format of the electronic health record. Natural language processing has been designed to convert free text into machine readable data. The aim of the current study was to ascertain the feasibility of using natural language processing to extract clinical information from >76,000 breast pathology reports. Approach and Procedure: Breast pathology reports from three institutions were analyzed using natural language processing software (Clearforest, Waltham, MA) to extract information on a variety of pathologic diagnoses of interest. Data tables were created from the extracted information according to date of surgery, side of surgery, and medical record number. The variety of ways in which each diagnosis could be represented was recorded, as a means of demonstrating the complexity of machine interpretation of free text. Results: There was widespread variation in how pathologists reported common pathologic diagnoses. We report, for example, 124 ways of saying invasive ductal carcinoma and 95 ways of saying invasive lobular carcinoma. There were >4000 ways of saying invasive ductal carcinoma was not present. Natural language processor sensitivity and specificity were 99.1% and 96.5% when compared to expert human coders. Conclusion: We have demonstrated how a large body of free text medical information such as seen in breast pathology reports, can be converted to a machine readable format using natural language processing, and described the inherent complexities of the task.

A pilot study evaluating shaved cavity margins with micro-computed tomography: a novel method for predicting lumpectomy margin status intraoperatively.

Microscopically clear lumpectomy margins are essential in breast conservation, as involved margins increase local recurrence. Currently, 18–50% of lumpectomies have close or positive margins that require re‐excision. We assessed the ability of micro‐computed tomography (micro‐CT) to evaluate lumpectomy shaved cavity margins (SCM) intraoperatively to determine if this technology could rapidly identify margin involvement by tumor and reduce re‐excision rates. Twenty‐five SCM from six lumpectomies were evaluated with a Skyscan 1173 table top micro‐CT scanner (Skyscan, Belgium). Micro‐CT results were compared to histopathological results. We scanned three SCM at once with a 7‐minute scanning protocol, and studied a total of 25 SCM from six lumpectomies. Images of the SCM were evaluated for radiographic signs of breast cancer including clustered microcalcifications and spiculated masses. SCM were negative by micro‐CT in 19/25 (76%) and negative (≥2 mm) by histopathology in 19/25 (76%). Margin status by micro‐CT was concordant with histopathology in 23/25 (92%). Micro‐CT overestimated margin involvement in 1/25 and underestimated margin involvement in 1/25. Micro‐CT had an 83.3% positive predictive value, a 94.7% negative predictive value, 83.3% sensitivity, and 94.7% specificity for evaluation of SCM. Evaluation of SCM by micro‐CT is an accurate and promising method of intraoperative margin assessment in breast cancer patients. The scanning time required is short enough to permit real‐time feedback to the operating surgeon, allowing immediate directed re‐excision.

Intraoperative micro-computed tomography (micro-CT): a novel method for determination of primary tumour dimensions in breast cancer specimens

OBJECTIVES Micro-CT is a promising modality to determine breast tumour size in three dimensions in intact lumpectomy specimens. We compared the accuracy of tumour size measurements using specimen micro-CT with measurements using multimodality pre-operative imaging. METHODS A tabletop micro-CT was used to image breast lumpectomy specimens. The largest tumour dimension on three-dimensional reconstructed micro-CT images of the specimen was compared with the measurements determined by pre-operative mammography, ultrasound and MRI. The largest dimension of pathologic invasive cancer size was used as the gold standard reference to assess the accuracy of imaging assessments. RESULTS 50 invasive breast cancer specimens in 50 patients had micro-CT imaging. 42 were invasive ductal carcinoma, 6 were invasive lobular carcinoma and 2 were other invasive cancer. Median patient age was 63 years (range 33-82 years). When compared with the largest pathologic tumour dimension, micro-CT measurements had the best correlation coefficient (r = 0.82, p < 0.001) followed by MRI (r = 0.78, p < 0.001), ultrasound (r = 0.61, p < 0.001) and mammography (r = 0.40, p < 0.01). When compared with pre-operative modalities, micro-CT had the best correlation coefficient (r = 0.86, p < 0.001) with MRI, followed by ultrasound (r = 0.60, p < 0.001) and mammography (r = 0.54, p < 0.001). Overall, mammography and ultrasound tended to underestimate the largest tumour dimension, while MRI and micro-CT overestimated the largest tumour dimension more frequently. CONCLUSION Micro-CT is a potentially useful tool for accurate assessment of tumour dimensions within a lumpectomy specimen. Future studies need to be carried out to see if this technology could have a role in margin assessment. ADVANCES IN KNOWLEDGE Micro-CT is a promising new technique which could potentially be used for rapid assessment of breast cancer dimensions in an intact lumpectomy specimen in order to guide surgical excision.

Using machine learning to parse breast pathology reports

Purpose Extracting information from electronic medical record is a time-consuming and expensive process when done manually. Rule-based and machine learning techniques are two approaches to solving this problem. In this study, we trained a machine learning model on pathology reports to extract pertinent tumor characteristics, which enabled us to create a large database of attribute searchable pathology reports. This database can be used to identify cohorts of patients with characteristics of interest.Methods We collected a total of 91,505 breast pathology reports from three Partners hospitals: Massachusetts General Hospital, Brigham and Women’s Hospital, and Newton-Wellesley Hospital, covering the period from 1978 to 2016. We trained our system with annotations from two datasets, consisting of 6295 and 10,841 manually annotated reports. The system extracts 20 separate categories of information, including atypia types and various tumor characteristics such as receptors. We also report a learning curve analysis to show how much annotation our model needs to perform reasonably.Results The model accuracy was tested on 500 reports that did not overlap with the training set. The model achieved accuracy of 90% for correctly parsing all carcinoma and atypia categories for a given patient. The average accuracy for individual categories was 97%. Using this classifier, we created a database of 91,505 parsed pathology reports.ConclusionsOur learning curve analysis shows that the model can achieve reasonable results even when trained on a few annotations. We developed a user-friendly interface to the database that allows physicians to easily identify patients with target characteristics and export the matching cohort. This model has the potential to reduce the effort required for analyzing large amounts of data from medical records, and to minimize the cost and time required to glean scientific insight from these data.

Breast Cancer Risk and Follow-up Recommendations for Young Women Diagnosed with Atypical Hyperplasia and Lobular Carcinoma In Situ (LCIS).

BackgroundThe risk of breast cancer in young women diagnosed with atypical hyperplasia and (LCIS) is not well defined. The objectives were to evaluate outcomes and to help determine guidelines for follow-up in this population.MethodsA retrospective review of women under age 35 diagnosed with ADH, ALH, LCIS, and severe ADH from 1987 to 2010 was performed. Patient characteristics, pathology and follow-up were determined from chart review.ResultsWe identified 58 young women with atypical breast lesions. Median age at diagnosis was 31 years (range 19–34). 34 patients had ADH, 11 had ALH, 8 had LCIS, and 5 had severe ADH.7 (12%) patients developed breast cancer. The median follow-up was 86 months (range 1–298). Median time to cancer diagnosis was 90 months (range 37–231). 4 cancers were on the same side, 3 were contralateral. 4 were IDC, 1 was ILC, and 2 were DCIS.Cancer was detected by screening mammogram in 4 patients, 2 by clinical exam, and 1 unknown. In the entire cohort, 26 (45%) patients had screening mammograms as part of their follow up, 12 patients had only clinical follow up, and 20 had no additional follow up. 13 patients required subsequent biopsies.ConclusionYoung women with atypical breast lesions are at a markedly increased risk for developing breast cancer and should be followed closely. Based on our findings, we recommend close clinical follow-up, MRI starting at age 25 through age 29, and screening mammograms for those over 30 in this high-risk group of patients.

Imaging of the Complications of Acute Pancreatitis

Epidemiology Acute pancreatitis is associated with a wide variety of complications [4–6] (Table 1). Negative prognostic factors that are associated with an increased likelihood of the development of complications in patients with acute pancreatitis include increasing age, gallstone disease, organ failure on admission, and pancreatic necrosis [5] (Fig. 1). Pancreatic necrosis is considered to be one of the most important complications that can occur, and it is also considered to be the most important indicator of disease severity [7]. Necrosis occurs in 6–20% of patients with acute pancreatitis, with 90% of affected patients developing at least single-organ failure and one third developing an infection [5]. Patients with pancreatitis without necrosis have a 6% morbidity rate and no mortality, but there is an 82% incidence of morbidity and a 23% incidence of mortality in patients with acute pancreatitis complicated by necrosis [1].

A Study of the Growth Patterns of Breast Carcinoma Using 3D Reconstruction: A Pilot Study

Lumpectomy with microscopically clear margins is a safe and effective approach for surgical management of breast carcinoma. Margins are positive for tumor in 18–50% of lumpectomies, as it is not possible to accurately determine the shape or microscopic borders of a tumor preoperatively or intraoperatively. We examined the 3D microanatomy and growth patterns of common breast carcinoma subtypes to provide guidance for lumpectomy surgery. Prospective consent was obtained for the use of excess tissue from patients undergoing lumpectomy or mastectomy for breast carcinoma. Tissue blocks from nine breast carcinomas were serially sectioned. Hematoxylin and eosin‐stained slides at 100 μm intervals were scanned using a Nanozoomer (Hamamatsu, Japan) microscopic‐resolution scanner. Three‐dimensional reconstructions of tumors were created from scanned images using Reconstruct, open‐access software. Breast carcinoma subtypes demonstrated characteristic growth patterns within breast tissue, which may have implications for lumpectomy surgery. Invasive ductal carcinomas showed a spherical shape, with a spiculated surface representing tumor cells infiltrating into surrounding parenchyma. Ductal carcinoma in situ appeared to spread along the duct system, creating dilated, tortuous, tumor‐filled ducts. The invasive lobular carcinomas examined had a haphazard, linear, infiltrative growth pattern, different from the shape seen in ductal carcinomas. Our preliminary work suggests that invasive ductal and invasive lobular carcinomas appear to have distinct growth patterns in three dimensions and ductal carcinoma in situ appears to grow in a linear fashion along the duct network. The microanatomy studies described have the potential to guide refinements in breast lumpectomy technique.

Aromatase inhibition to decrease background parenchymal enhancement: premedication before magnetic resonance imaging?

M agnetic resonance imaging (MRI) of the breast is a useful tool in diagnosing mammographically occult cancers and can provide, in selected circumstances, additional diagnostic information on women with newly diagnosed breast cancer. At least one prospective trial supports the use of breast MRI as an adjunct to mammography in screening women at high risk of breast cancer, although no survival data are available as of yet. Although the sensitivity of breast MRI approaches 90%, the specificity remains low, with quoted percentages of 37% to 98%. Nonspecific enhancement of normal breast parenchyma can lead to a high rate of call backs for additional diagnostic tests and an increased biopsy rate in many women. Patients considered Bhigh risk[ and therefore most likely to benefit from MRI, despite a high risk of a false-positive result, have been defined by the American Cancer Society guidelines for MRI screening, which limit screening MRI to women with a known BRCA mutation, firstdegree relatives of BRCA mutation carriers, and women with a lifetime risk of breast cancer higher than 20% to 25% as estimated by risk assessment models, such as the BRCAPRO, a statistical model and software which predicts the genetic predisposition to breast and ovarian cancer. Rather than use the risk models to estimate lifetime risk, we prefer genetic testing for BRCA mutations in high-risk women (those with a 910% risk of carrying a mutation). Murphy et al, in the analysis of the BRCA PRO results of 18,190 women undergoing screening mammography, found that a significant number of women who were predicted mutation carriers would not meet the American Cancer Society criterion of greater than 20% lifetime risk and thus would not be offered breast MRI. To comprehensively identify the MRI target population, genetic testing should be encouraged in all women identified as high risk, and MRI should be considered in those who test positive or indeterminate. In the current issue of Menopause, Mousa et al attempt to address the issue of the high rate of false positives associated with MRI. The authors present a novel approach aimed to increase the specificity of breast MRI by pretreating a cohort of women with the aromatase inhibitor (AI) letrozole to decrease background parenchymal enhancement. Previous studies have demonstrated the effects of physiologic hormonal status, hormone therapy, and selective estrogen receptor modulators on the background enhancement on breast MRI. AIs block estrogen synthesis in peripheral tissues, which is the primary source of estrogen in postmenopausal women and have become the treatment of choice for postmenopausal patients with endocrine-sensitive breast cancer, resulting in significantly improved disease-free survival. The only other study examining the effects of AIs on MRI breast parenchymal enhancement to date has been a retrospective review of pretreatment and posttreatment breast MRIs of 168 women taking AIs for breast cancer by Goldfarb et al, presented in poster form at the 2011 San Antonio Breast Cancer Symposium. The authors found that adjuvant AIs were associated with a significant decrease in breast parenchymal enhancement on MRI. The current study examines the effect of a short course of high-dose letrozole on breast parenchymal enhancement in postmenopausal women. The dose of letrozole was arbitrarily chosen and is significantly higher than that used in current clinical practice, although well below the maximum tolerated dose. The 3-day duration of treatment seemed to be well tolerated, with no serious events noted, although concurrent use of hormone therapy (HT) by 12 of the 14 study participants may have lowered the incidence and severity of the adverse effects reported. It is acknowledged by the authors that future studies including women not taking HT would be of value. The study population comprised 14 postmenopausal women with an average age of 60 years. Postmenopausal women would not be a population who would routinely undergo breast MRI, unless they fell into a high-risk group as defined in the opening paragraph or were undergoing MRI as part of the diagnostic workup for newly diagnosed breast cancer. Twelve of 14 women were taking HT, which has previously been shown to increase background enhancement on breast MRI. The authors state that because HT was constant for the duration of the study, any decrease in background enhancement was caused solely by the effect of letrozole. It is also worth noting that although background enhancement was decreased in 10 of 14 patients, it was significantly decreased in only half (7 of 14) of the study population. Although the study demonstrated a measurable reduction in breast parenchymal enhancement when patients were pretreated with a short course of letrozole, it must also be considered that a global reduction in parenchymal enhancement may obscure a malignant lesion. Indeed, this is pointed out by the authors in their discussion. The approach described is interesting, however, it is not ready for prime time, and we do not recommend this approach

Doctor, Would You Drain This Collection?

3 A 53-year-old man with cholangiocarcinoma (arrowhead, Figure, a) presented for computed tomography (CT) before commencing proton-beam therapy. Two thick-walled fluidcontaining perihepatic lesions were seen, traversed by rippling lines of air density, between the liver and gastroesophageal junction (curved arrow, Figure, a), and duodenum (straight arrow, Figure, a; arrow, Figure, b). These were deemed to have been caused by AlloDerm (LifeCell, Branchburg, New Jersey), an implantable material that had been placed since the earlier CT scan depicting the cholangiocarcinoma (arrowhead,