Automatic structuring of radiology free-text reports.

A natural language processor was developed that automatically structures the important medical information (eg, the existence, properties, location, and diagnostic interpretation of findings) contained in a radiology free-text document as a formal information model that can be interpreted by a computer program. The input to the system is a free-text report from a radiologic study. The system requires no reporting style changes on the part of the radiologist. Statistical and machine learning methods are used extensively throughout the system. A graphical user interface has been developed that allows the creation of hand-tagged training examples. Various aspects of the difficult problem of implementing an automated structured reporting system have been addressed, and the relevant technology is progressing well. Extensible Markup Language is emerging as the preferred syntactic standard for representing and distributing these structured reports within a clinical environment. Early successes hold out hope that similar statistically based models of language will allow deep understanding of textual reports. The success of these statistical methods will depend on the availability of large numbers of high-quality training examples for each radiologic subdomain. The acceptability of automated structured reporting systems will ultimately depend on the results of comprehensive evaluations.     

Reliability of the international knee documentation committee radiographic grading system

BACKGROUND: The International Knee Documentation Committee (IKDC) forms are commonly used to measure outcomes after anterior cruciate ligament (ACL) reconstruction. The knee examination portion of the IKDC forms includes a radiographic grading system to grade degenerative changes. The interrater and intrarater reliability of this radiographic grading system remain unknown. HYPOTHESIS: We hypothesize that the IKDC radiographic grading system will have acceptable interrater and intrarater reliability. STUDY DESIGN: Case series (diagnosis); Level of evidence, 4. METHODS: Radiographs of 205 ACL-reconstructed knees were obtained at 5-year follow-up. Specifically, weightbearing posteroanterior radiographs of the operative knee in 35 degrees to 45 degrees of flexion and a lateral radiograph in 30 degrees of flexion were used. The radiographs were independently graded by 2 sports medicine fellowship-trained orthopaedic surgeons using the IKDC 2000 standard instructions. One surgeon graded the same radiographs 6 months apart, blinded to patient and prior IKDC grades. The percentage agreement was calculated for each of the 5 knee compartments as defined by the IKDC. Interrater reliability was evaluated using the intraclass correlation coefficient (ICC) 2-way mixed effect model with absolute agreement. The Spearman rank-order correlation coefficient (r(s)) was applied to evaluate intrarater reliability. RESULTS: The interrater agreement between the 2 surgeons was 59% for the medial joint space (ICC = 0.46; 95% confidence interval [CI] = 0.35-0.56), 54% for the lateral joint space (ICC = 0.45; 95% CI = 0.27-0.58), 49% for the patellofemoral joint (ICC = 0.40; 95% CI = 0.26-0.52), 63% for the anterior joint space (ICC = 0.20; 95% CI = 0.05-0.34), and 44% for the posterior joint space (ICC = 0.28; 95% CI = 0.15-0.40). The intrarater agreement was 83% for the medial joint space (r(s) = .77, P < .001), 86% for the lateral joint space (r(s) = .76, P < .001), 81% for the patellofemoral joint (r(s) = .79, P < .001), 91% for the anterior joint space (r(s) = .48, P < .001), and 69% for the posterior joint space (r(s) = .64, P < .001). CONCLUSIONS: While intrarater reliability was acceptable, interrater reliability was poor. These findings suggest that multiple raters may score the same radiographs differently using the IKDC radiographic grading system. The use of a single rater to grade all radiographs when using the IKDC radiographic grading system maximizes reliability.     

A patient documentation system for radiation therapy by computer-planning. 2. 5-year experience with the Patient Documentation System PATDOK and recording of irradiation data

About a five years experience with our data base system PATDOK, in which are stored now more than 5000 patients, is reported. Further a method for coding treatment regimes is proposed. First results are discussed.     

Experience with breast biopsies using the Advanced Breast Biopsy Instrumentation system

OBJECTIVE: A retrospective review of our experience with advanced breast biopsy instrumentation (ABBI) was undertaken to evaluate its efficacy for excisional breast biopsy of lesions detected on mammography. To our knowledge, experience with ABBI has not been previously reported in the radiology literature. MATERIALS AND METHODS: Biopsies using the ABBI system and an adapted dedicated table were performed in 53 patients who had 54 mammographically evident lesions. Samples were obtained with cannulas ranging in size between 5 and 20 mm. Indications for biopsy were calcifications (n = 22) and masses (n = 31). RESULTS: Forty-five specimens (44 patients) had benign results at pathology: 15 specimens were diagnosed as fibroadenoma, 15 as cystic breast disease, and four as reactive lymph nodes; the remaining 11 specimens had benign diagnoses of adenosis, fibrosis, and hyperplasia. The average specimen size was 4.8 cm in greatest longitudinal dimension. One patient had a nondiagnostic biopsy for calcifications later found to be dermal. Seven patients were diagnosed as having breast cancer; in six of these, the tumor involved the margins of the specimen. One patient had marked atypia that required reexcision for the diagnosis of intraductal carcinoma to be made. CONCLUSION: The ABBI procedure is a more invasive and less readily available procedure than core needle biopsy for sampling of benign breast lesions. In seven cases of malignancy diagnosed at breast biopsy, the ABBI technique completely excised only a single lesion. In our experience, the ABBI procedure offered no advantages over core needle biopsy for either benign or malignant lesions.     

Experience in implementing the Mark III Rad 8 system for radiotherapy dose computations.

Rad 8 is a small, in-house computer system for calculating radiotherapy dose distributions, both for multiple field and rotational external beam techniques and for interstitial and intracavitary redium treatments. Implementing the system has its difficulties, particularly in obtaining the necessary beam profile information. Several approaches have been tried and finally a separate profile generator program was written. Other difficulties encountered with beam data storage were overcome by making minor changes to the system programs. The various calculation techniques have been evaluated, and in general found to be satisfactory, and suitable factors to be used in the inhomogeneity correction for 4 MV X rays have been calculated based on previously published measured data. An option to change beam weightings exists and this can be used to renormalize dose distributions to give percentage isodose curves relative to the prescribed tumour dose. If this procedure is followed dose homogeneity over the target volume can readily be assessed, and the exposure times may easily be calculated from the new weightings. Certain shortcomings in the rotating beam program and the radium program may easily lead to error if extreme care is not exercised.     

New directive for documentation review in radiation oncology.

A new directive from the Health Care Financing Administration regarding reimbursement for daily patient care underscores the importance of accurate documentation in radiation oncology. Ms. Blue describes a quality control system in use at her institution which involves a comparison of what is documented in the medical record with what has been charged.     

Kaposi's sarcoma. Experience with ten cases

Kaposi's sarcoma is uncommon in the U.S.A. and common in Africa. Men are affected at least 4 times more frequently than females. It occurs simultaneously with lymphomas and resembles them in several ways. Radiation therapy controls the disease well. Ten patients were seen with the disease. Most had a relatively benign course, especially the 4 women. Treatment with supervoltage irradiation and cryosurgery was effective in controlling local disease.