William H. Wolberg

Breast Cancer Diagnosis and Prognosis Via Linear Programming

Two medical applications of linear programming are described in this paper. Specifically, linear programming-based machine learning techniques are used to increase the accuracy and objectivity of breast cancer diagnosis and prognosis. The first application to breast cancer diagnosis utilizes characteristics of individual cells, obtained from a minimally invasive fine needle aspirate, to discriminate benign from malignant breast lumps. This allows an accurate diagnosis without the need for a surgical biopsy. The diagnostic system in current operation at University of Wisconsin Hospitals was trained on samples from 569 patients and has had 100% chronological correctness in diagnosing 131 subsequent patients. The second application, recently put into clinical practice, is a method that constructs a surface that predicts when breast cancer is likely to recur in patients that have had their cancers excised. This gives the physician and the patient better information with which to plan treatment, and may eliminate the need for a prognostic surgical procedure. The novel feature of the predictive approach is the ability to handle cases for which cancer has not recurred (censored data) as well as cases for which cancer has recurred at a specific time. The prognostic system has an expected error of 13.9 to 18.3 months, which is better than prognosis correctness by other available techniques.

Development and Pilot Evaluation of a Computer-Based Support System for Women with Breast Cancer

A computer-based support system was developed to help women cope with the crisis of breast cancer. The system, called Comprehensive Health Enhancement Support System (CHESS)contains integrated information, referral, decision, and social support programs. It was developed with intensive input from potential users through needs-assessment surveys and field testing. This article reports on the results of two pilot studies involving 30 women with breast cancer. The preliminary versions of CHESS were used extensively by older and younger women and by college and high school graduates. Participants in the pilot studies suggested several content enhancements and user-friendly aids for the developing system. User surveys indicated that CHESS was easy to use and would be valuable to other women with breast cancer, their partners, and their adult children. The women reported that they experienced more positive emotions and fewer negative emotions as a result of using the system.

Machine learning techniques to diagnose breast cancer from image-processed nuclear features of fine needle aspirates

An interactive computer system evaluates and diagnoses based on cytologic features derived directly from a digital scan of fine-needle aspirate (FNA) slides. A consecutive series of 569 patients provided the data to develop the system and an additional 54 consecutive, new patients provided samples to test the system. The projected prospective accuracy of the system estimated by tenfold cross validation was 97%. The actual accuracy on 54 new samples (36 benign, 1 atypia, and 17 malignant) was 100%. Digital image analysis coupled with machine learning techniques will improve diagnostic accuracy of breast fine needle aspirates.

Psychosexual adaptation to breast cancer surgery

Disturbances associated with a breast cancer diagnosis were defined when psychological assessments from 63 patients with a known breast cancer diagnosis were compared to those from 56 patients with an as yet undiagnosed malignancy. Subsequent assessments from the 56 patients with an undiagnosed breast cancer showed disturbances after they saw a physician compared with the assessments from 72 similar patients ultimately diagnosed as benign. Apprehension apparently arose from clues given before a biopsy was done even though the cancer was not yet diagnosed. Compared with the benign breast disease group, the disturbances in patients suspected or diagnosed with breast cancer were found chiefly in assessments of mood and adjustment, and less in assessments of more durable characteristics of personality, psychopathology, and sexual behavior. Psychological problems associated with breast cancer decreased over time, but residuals persisted for at least 16 months postoperatively. Few differences were found between 41 patients who elected breast‐conserving surgery and 78 who were treated with mastectomy. Problems were not eliminated by operations which saved the breast.

The study of burn wound edema using dichromatic absorptiometry.

Burn wound edema is a source of major morbidity and mortality. To quantitatively study this edema, we have devised a noninvasive method called Dichromatic Absorptiometry (DA) which is precise (2-3%) and accurate (r equal to 0.99) in measuring changes in tissue fluid. A scanning technique using a dual photon source, 125I and 241Am, is used. The variable attenuation of the two photon energies allows for selective mass measurements of fluid, protein, lipid, and bone. In limbs with thermal injury the correlation of DA changes in fluid mass with circumference was (r equal to 0.97), but DA was more sensitive. In deep second- and third-degree burns, edema formation was maximum between 12 and 18 hours postburn with 80% of maximum already present at 4 hours. Resorption was complete by 1 week in second-degree burns, but significantly delayed in third-degree burns. Superimposed infection produced persistent edema.

The Effect of Immediate and Delayed Cold Immersion on Burn Edema Formation and Resorption

Massive wound edema after a burn may impair healing and help to convert partial to full-thickness injury. Cold treatment (usually by immersion) has been reported to decrease wound edema and is useful in first-aid treatment of burns. Reliable quantitative data have been lacking and frequently a superficial burn has been studied. Since cold by decreasing peripheral blood flow could actually be harmful to a deep burn, especially if applied late, we measured the effect of cold immediately and 2 minutes postburn on edema formation and resorption in a deep second-degree burn in sheep hindlimbs. We used Dichromatic Absorptiometry, a noninvasive, reliable method for measuring tissue fluid, to quantitate edema. Immediate application of cold by immersion in 15 degrees C saline for 30 minutes reduced the edema of a deep second-degree burn and did not impair resorption rate compared with control limbs, fluid content returning to baseline after 1 week. Cold treatment beginning 2 minutes after the burn did not decrease edema formation and did impair resorption. Fifteen per cent of the edema fluid was still present 1 week postburn, suggesting further injury to the burn wound vasculature with use of cold immersion 2 minutes postburn.

Factors influencing options in primary breast cancer treatment.

Primary breast cancer treatment is determined by tumor factors and by patient preference. Breast cancer treatments that preserve the cosmetic appearance of the breast are appealing and effective for appropriately selected patients; long-term survival following tumor excision and breast irradiation appears to be comparable to that for mastectomy. Since April 1981, when a protocol was developed and treatment options were offered, factors influencing treatment selection have been analyzed in 206 consecutive primary breast cancer patients. Mastectomy was dictated by tumor-related factors in 96 patients (47%); 110 patients (53%) had the option of mastectomy or conservation--tumor excision plus radiotherapy to the breast. Among these 110 eligible patients, 54 chose conservation (49%) and 56 chose mastectomy (51%). Intraoperative findings for ten patients electing conservation necessitated mastectomy, so conservation was accomplished for 44 (21%) of those treated for breast cancer. Beginning in July 1982, breast cancer patients took a battery of psychosexual assessments before any operation (Profile of Mood States [POMS], Health Locus of Control Scale [HLCS] Locke-Wallace Marital Adjustment Test [MAT], Psychosocial Adjustment to Illness Scale [PAIS], Derogatis Sexual Function Inventory [DSFI], Millon Clinical Multiaxial Inventory [MCMI], and a Breast Cancer Information Test [BCIT]). Comparisons of psychologic and demographic variables were made between patients who chose mastectomy and those who chose conservation. No demographic variable was statistically significantly related to choice, although older women tended to select mastectomy more than younger women. Compared with those who elected conservation, women who elected mastectomy were more tense and anxious (P less than .01), more introverted (P less than .01), felt more depressed and dejected (P less than .05), and reported more sexual problems (P less than .05). Those who elected conservation valued their physical appearance more highly (P less than .01) and were generally more self-interested (P less than .05). Mastectomy was dictated by medical considerations for approximately half of patients with breast cancer. Among candidates for breast conservation, the importance of retaining the breast appeared to be determined to a significant degree by measurable psychological factors.

Survival-Time Classification of Breast Cancer Patients

The identification of breast cancer patients for whom chemotherapy could prolong survival time is treated here as a data mining problem. This identification is achieved by clustering 253 breast cancer patients into three prognostic groups: Good, Poor and Intermediate. Each of the three groups has a significantly distinct Kaplan-Meier survival curve. Of particular significance is the Intermediate group, because patients with chemotherapy in this group do better than those without chemotherapy in the same group. This is the reverse case to that of the overall population of 253 patients for which patients undergoing chemotherapy have worse survival than those who do not. We also prescribe a procedure that utilizes three nonlinear smooth support vector machines (SSVMs) for classifying breast cancer patients into the three above prognostic groups. These results suggest that the patients in the Good group should not receive chemotherapy while those in the Intermediate group should receive chemotherapy based on our survival curve analysis. To our knowledge this is the first instance of a classifiable group of breast cancer patients for which chemotherapy can possibly enhance survival.

Gastric ulceration in patients receiving intrahepatic infusion of 5-fluorouracil.

Gastric ulceration developed in eight patients during intrahepatic arterial infusion of 5-FU. Bleeding occurred in four instances and perforation in one. In all cases the catheter tip had been dislodged and was proximal to its correct position, allowing the stomach to be directly infused with 5-FU. No duodenal ulcers were noted. All patients were symptomatic for several days before the diagnosis was made. Of 20 patients with catheter dislodgement, five had documented ulcers, three had upper gastrointestinal bleeding of undetermined etiology, eight had epigastric pain or vomiting and only four were asymptomatic. Prompt determination of catheter position is necessary in patients receiving intrahepatic arterial infusion of 5-FU if symptoms consistent with gastric ulceration occur. Gastric ulcers should be vigorously treated because of the high rate of complications in patients receiving chemotherapy.

A comparison of methods for the production of monodispersed cell suspensions from human primary breast carcinomas

SummaryProduction of monodispersed cell suspensions from primary human breast tumors is difficult due to the predominant stromal composition of most breast tumors. Our studies were designed to optimize dispersion of breast tumors of known stromal content and histopathology. In a first series of experiments three enzymatic protocols were compared to disperse minced tissue: (A) treatment with collagenase (2 mg/ml) in the presence of 5% serum for 24 hours; (B) treatment with collagenase (6 mg/ml) and DNase (0.002%) in 10% serum for 3 hours; (C) treatment with collagenase (2 mg/ml) for 3 hours followed by pronase (0.075%) for l hour. Protocol A produced better cell yields than B or C for all tumors tested. The monodispersed cells were suspended in a 0.3% semi-solid agar with alpha modified Eagles medium (αMEM), 10% serum, and selected hormones, then layered over similarly enriched 0.5% semi-solid agar. The cells prepared by protocol A had a higher plating efficiency and larger average colony size than B or C. In a second series of experiments, protocol A was repeated and compared to two additional procedures: (D) treatment with collagenase (2 mg/ml) and hyaluronidase (1 mg/ml) in the presence of 5% serum for 24 hours; and (E) mechanical disaggregation. Protocol D exhibited a small but significant negative difference from A, while E was the least efficient in producing viable monodispersed cells from the tumors. All enzymatically monodispersed cells produced clonal growth in our agar system. However, mechanically dispersed cells gave growth in only 4 of 7 tumors. Protocol A, in addition to yielding the highest number of viable cells per gram of tissue, gave the highest plating efficiency of all protocols tested.