Thomas R. TenHave

The impact of 2D versus 3D quantitation of tumor bulk determination on current methods of assessing response to treatment

PURPOSEnMeasurements from sequential axial 2D data in cancer patients are commonly used to assess treatment response or disease progression. This study compares the volume of tumor bulk calculated with 3D reconstructions with that calculated by conventional methods to determine if it might change patient classification.nnnMETHODnAll medical, gynecologic, and pediatric oncology patients under treatment who were evaluated with serial CT scans between January 1, 1992, and July 31, 1994, for whom the digital data were available were included in this study. For each tumor site, the maximum diameter and its perpendicular were measured and multiplied together to yield an area. The sum of areas of the measured lesions was used as an approximation of overall 2D tumor volume. In addition, the 2D area of each site was multiplied by its height, yielding a 2D volume. Last, the digital data were loaded into a 3D computer system and total 3D tumor volumes determined. All medical and gynecologic oncology patients were treated based upon the 2D area of tumor. The pediatric oncology patients were treated based upon the 2D volume of tumor measured as per standard practice. The members of each treating oncologic service assessed their patients as to how the other two methods would have changed their classification of the patients response category.nnnRESULTSnFour hundred thirty-three CT scans were performed in 139 patients, which included 204 baseline and 294 follow-up CT examinations. Seventy patients had new tumor foci and would have been classified as failure by all three methods of tumor bulk measurement. The 3D volume versus the 2D area method of tumor bulk assessment would have changed response categories in 52 of the 294 follow-up CT examinations (p < 0.0001). Thirty-five patients were recategorized from either no response to failure (21 patients) or no response to response (14 patients) categories. If only those follow-up studies without new metastatic foci are considered, the 3D volume versus the 2D area methods of tumor assessment would have changed the treatment response category in 23.2%. The use of the 2D volume method of calculating tumor volume of bulk tended to overestimate the overall tumor size by an average of 244 cm3 (p = 0.001).nnnCONCLUSIONnThe 3D method of tumor volume measurement differs significantly from conventional 2D methods of tumor volume determination. Large prospective studies analyzing the usefulness of 3D tumor volume measurements and assessing possible changes in patient response categories would be required for full utilization of this more accurate method of following disease bulk.

Comparing Hospital Mortality in Adult Patients with Pneumonia: A Case Study of Statistical Methods in a Managed Care Program

Comparing hospitals by their outcomes has become a popular method of drawing inferences about their relative quality. Outcome studies are now done regularly by employer- and insurer-sponsored managed care programs [1] and by federal [2] and state agencies [3-5]. Mortality, a well-defined outcome, has received the most study, and the public availability of raw mortality data has led to their frequent use as a measure of quality of care. Differences in mortality are known to be attributable to differences in patients severity of illness [6, 7] or comorbid conditions [8]; thus, researchers have tried to develop methods to adjust raw data for these factors [9]. A well-described danger of such methods lies in our inability to identify all clinical predictors of mortality from computerized data sets [10-13]. In addition, a fundamental dispute remains about whether there is any correspondence between mortality and quality of care [14, 15]. The desire of health care payers for simple measures of hospital performance conflicts with the skepticism of health care providers about the validity and fairness of the assessment process [16, 17]. A less frequently discussed but equally vital issue in the comparison of patient outcomes is the appropriateness of the statistical methods used for analyzing data and interpreting results. Raw data are clearly insufficient for comparing hospitals or physicians; some statistical methods must be invoked. Ideally, using a set of techniques called standardization [18], the comparison of different samples of patients in varied clinical settings should be feasible. We examine one attempt to compare hospitals based on their observed mortality; this attempt was made in 1992 by consultants to a managed care program for a large corporation. The corporation sought to determine which of the hospitals serving the corporations employees in central Pennsylvania delivered better quality of care as reflected in part by fewer in-hospital deaths in 1989 and 1990. Partly on the basis of the methods reported here, the corporation selected 10 of these hospitals to be eligible for its managed care network. Focusing on adult pneumonia, which was one of the diagnostic groups used by the consultants, we describe the consultants methods and re-create their results for 1989 and 1990. Then, we reassess those results using more appropriate methods, and finally, we validate results using newly acquired 1991 data. (A glossary of statistical terms is given in Appendix 1.) Methods Choice of Diagnosis-Related Groups The study sample included data reported to the Pennsylvania Health Care Cost Containment Council on all inpatient hospitalizations for adult pneumonia (diagnosis-related groups 089-090) in 1989 and 1990, and, later, in 1991, for 22 hospitals in central Pennsylvania. Because we were using the consultants methods, we excluded from analysis all patients older than 65 years of age, for whom Medicare would likely be the primary insurer. In addition, we excluded four patients with the acquired immunodeficiency syndrome and seven awaiting organ transplantation because only one hospital provided these specialized kinds of care. Examination of the 1991 data did not begin until after completion of the re-analysis for the years 1989 and 1990. The public-use data sets include coded discharge abstracts and the MedisGroups (Mediqual Systems, Westborough, Massachusetts) admission severity group, an indicator of severity of illness [19]. Admission severity group attempts to distill many key clinical findings at patient admission into a single, five-level severity score. In the years 1989-1991, admission severity group was not diagnosis specific; a patient received a given admission severity group score regardless of the cause of his or her illness. All hospitals in Pennsylvania must subscribe to the MedisGroups system, calculate admission severity groups retrospectively from medical records, and report these data and computerized discharge abstracts to the Cost Containment Council. The Managed Care Consultants Analysis The consultants relied only on admission severity group to adjust for both severity of acute illness and comorbid conditions. Patient age, which is not an element in admission severity group, was not used. The consultants statistical technique is best described as a stratified, standard, normalized analysis. First, they calculated death rates for each admission severity group within each institution. Next, they computed the unweighted mean and standard deviations for admission severity group death rates for all hospitals. They then computed a standard, normalized [20] difference between the hospital and the average death rate for each level of admission severity group by dividing the difference by the standard deviation of death rates. Finally, they combined these normalized differences across diagnosis-related group levels within each hospital by a weighted average. Weights depended on the fraction of the hospitals admissions in each admission severity group. The consultants did not calculate standard errors or P values for these weighted averages of normalized differences; they displayed their results along a horizontal line, indicating the location of the average normalized difference and the relative position of a given hospital along the line. Further details of our implementation of the consultants methods are given in Appendix 2. Methods of Reanalysis Our reanalysis of the data on adult pneumonia from the same 22 hospitals sought to confirm or refute the consultants results with conventional methods for data with binary outcomes, in this case, death or survival at discharge. We implemented model-based standardization [21] to compare hospital mortality adjusted for multiple patient risk factors. First, we verified the association between death and the patients admission severity group by logistic regression as implemented in SAS, Proc Logistic (SAS Institute, Cary, North Carolina). Next, the hospital with the worst outcome in 1989-1990 (hospital 1) as calculated by the managed care consultants method Table 1, was compared with the others by fitting two nested regression models: one with only admission severity group as a covariate, the other with admission severity group plus an indicator variable to detect any difference between hospital 1 and all other area facilities. The likelihood ratio statistic for the differences between the two models showed whether the hospital with the worst outcome, according to the consultants analysis, was truly different from the other hospitals. Table 1. Mortality for Adults Patients with Pneumonia in 22 Central Pennsylvania Hospitals, 1989-1991* Also relevant to the outcome of many acute diseases is patient age, a factor readily available in the public-use data set but not used by the consultants. Based on a preliminary calculation showing that mortality increased with patient age, patients were grouped into three age ranges: 18 to 39 years, 40 to 54 years, and 55 to 64 years. For age to affect interhospital comparisons of pneumonia mortality, two conditions must be met: The distribution of ages must differ across hospitals and age must be associated with death after controlling for admission severity group. We tested the first condition by cross-classifying the three age categories and the 22 hospitals and calculating the Mantel-Haenszel chi-square statistic for differences in mean age across hospitals. We tested the second condition by using the likelihood ratio statistic for the difference in two logistic regression models: one with both admission severity group and age as factors and the other with only admission severity group as a factor. The last step in testing whether hospital 1 differed significantly from the others again involved the comparison of two nested models. For this comparison, the two models included both age and admission severity group as standardization factors, and one also included an indicator variable for hospital 1. The ratio of the odds of mortality and a P value were computed as previously described. After determining whether hospital 1 had an adjusted mortality higher than that of the other hospitals, we investigated whether another hospital among the 22 might have had significantly higher mortality than expected. Logistic regression also allows for the calculation of predicted probabilities of death for each patient with a given age and admission severity group level. By summing the number of deaths and this predicted probability of death over all patients with pneumonia within each hospital, we computed at the hospital level both an observed and a predicted (or expected) number of deaths. Hospital 4 showed the largest positive difference between observed and predicted numbers of deaths and thus became the test hospital for further analysis of differences in mortality. As before, we fit two nested regression models, one with admission severity group and age as covariates and the other with an additional single indicator variable for hospital 4, to test whether patient mortality was significantly higher there than at other facilities. Multiple Comparisons Contrasts among hospital death rates in the consultants study, as in typical outcomes analyses, were determined by the data rather than planned in advance. Each hospital was compared in turn to determine whether it differed from all others. In this situation, a critical P value of 0.05 used repeatedly for each hospital comparison did not function to limit type I error, which was the probability of falsely classifying a hospital as a mortality outlier. Diehr and colleagues [22, 23] have commented on the issue of multiple comparisons, or multiple statistical tests, in a related context. The Bonferroni method [24], one of several available techniques [25], was used to adjust the critical P value to control type I error for multiple comparisons of hospitals. Computer Simulatio

Literacy assessment in a cardiovascular nutrition education setting

We assessed functional literacy of hypercholesterolemic or hypertensive African Americans (n = 339) prior to their participation in a nutrition education program. A word pronunciation and recognition test using 20 common cardiovascular or nutrition terms was first developed based on correlations with standardized reading achievement test scores, then administered to program participants. Nearly half (48%) had word recognition scores equivalent to a < or = 8th grade reading level. Lower scores were associated with less education, lower income, unemployment, heavier work activity if employed, less healthy diets, history of heart disease or diabetes, and higher depression scores (all P < 0.01); several of these associations were independent of education. The educational materials were geared to a 5th to 8th grade reading level. However, when both audiotaped and printed instruction were provided, individuals with reading scores < or = 8th grade preferentially used the tapes. This brief and relatively unobtrusive literacy assessment may help to identify persons who can benefit most from audiovisual approaches to cardiovascular nutrition education.

The quality of 3D reconstructions from 1.0 and 1.5 pitch helical and conventional CT.

PURPOSEnCT data are commonly used to create 3D images. For this purpose, thin and overlapped slices are desirable. Helical (spiral) CT offers the ability to adjust the slice reconstruction interval from 0 to 100%. However, its use in 1.0 and 1.5 pitch helical CT and 3D imaging, especially with respect to surface detail, is relatively untested.nnnMETHODSnTen objects selected for their varying size, shape, and density were scanned (fourth generation Picker PQ2000) by contiguous 2,4 and 8 mm conventional and helical sequences. The latter were obtained with a pitch of both 1.0 and 1.5 and were reconstructed into a 3D image with 0-75% overlapping of the reconstructed slices. Each of the 24 different sequences per scanned object was reconstructed into identical sets (projections) of 3D images displayed on color film. The 24 3D image sets for each object were submitted to six blinded radiologists who separately ranked them from best to worst.nnnRESULTSn3D reconstructions obtained from CT scans with a thinner slice thickness, half-field (15 cm FOV), and helical technique were rated as statistically superior. The 1.0 and 1.5 helical sequences obtained with a 4 or 8 mm slice thickness scored statistically better than 3D reconstructions from equivalent conventional scans. Overlapping of the reconstructed helical slices by 25-75% generally improved the quality of the 3D reconstruction.nnnCONCLUSIONnHelical CT with either a 1.0 or a 1.5 pitch offers the ability to obtain higher quality 3D reconstructions than from comparable conventional CT scans.

Capillary versus aspiration biopsy: Effect of needle size and length on the cytopathological specimen quality

PurposeTo test the value of the nonaspiration, or capillary, biopsy technique by experimental comparison with the conventional fine-needle aspiration technique using various needle gauges and lengths.MethodsOn fresh hepatic and renal tissue from five autopsies, multiple biopsy specimens were taken with 20, 22, and 23-gauge Chiba needles of 5, 10, 15, and 20-cm length, using the aspiration technique and the capillary technique. The resultant specimens were graded on the basis of a grading scheme by a cytopathologist who was blinded to the biopsy technique.ResultsThe capillary technique obtained less background blood or clot which could obscure diagnostic tissue, although not significantly different from the aspiration technique (p=0.2). However, for the amount of cellular material obtained, retention of appropriate architecture, and mean score, the capillary technique performed statistically worse than aspiration biopsy (p<0.01). In addition, with decreasing needle caliber (increasing needle gauge) and increasing length, the capillary biopsy was inferior to the aspiration biopsy.ConclusionThe capillary biopsy technique is inferior to the aspiration technique according to our study. When the capillary technique is to be applied, preference should be given to larger caliber, shorter needles.

In Vivo Accuracy Of Two Radiographic Systems In The Detection Of Björk-Shiley Convexo-Concave Heart Valve Outlet Strut Single Leg Separations☆☆☆★★★♢

OBJECTIVEnModified cineradiographic systems have been used clinically to detect partially broken outlet struts in normally functioning Björk-Shiley convexo-concave heart valves. Almost all such valves were explanted, presuming that full failure would likely follow. Inasmuch as the clinical setting only rarely permits examination of normally rated valves, the accuracy of radiographic detection cannot be clinically defined. This study uses the clinical radiographic technique in sheep implanted with known-status convexo-concave valves, comparing its accuracy and that of a newly developed, geometric image magnification radiography system.nnnMETHODSnTwenty-one sheep with mitral convexo-concave valves were studied on both systems. Five were used for extensive training. When operators were expert with both systems, images of four intact valves and 12 valves with outlet strut single leg separations, along with a seventeenth single leg separation valve used for calibration, were integrated into 112 image sets organized into a balanced incomplete block design for evaluation by eight trained, blinded reviewers.nnnRESULTSnCineradiography sensitivity was 24% versus 31% for direct image magnification. The odds ratio for detection of single leg separation by direct image magnification versus cineradiography was 2.0 (95% confidence interval, 0.76 to 5.9; p = 0.13). Cineradiography specificity was 93% versus 90% for direct image magnification. Sensitivity and specificity varied markedly by reviewer, with sensitivity ranging from 8% to 55% and specificity from 51% to 100% for the combined technologies.nnnCONCLUSIONSnThe data support the need for more intensive training for convexo-concave valve imaging and further investigation of unconventional radiographic technologies. Clinical cineradiography of convexo-concave valves may detect as little as 25% of valves having a single leg separation, underestimating the prevalence of single leg separations and thereby implying more rapid progression to full fracture than is actually the case.

The clinical usefulness of routine stacked multiplanar reconstruction in helical abdominal computed tomography.

RATIONALE AND OBJECTIVESnThe authors evaluate the usefulness of stacked multiplanar reconstructions in routine, thick-section abdominal computed tomography.nnnMATERIALS AND METHODSnTwenty-five routine, thick-section contrast abdominal CTs performed with equivalent technique were reformatted by multiplanar reconstructions in sagittal and coronal planes sequentially from side-to-side and front-to-back. The image sets were submitted, first axial images only followed by axial plus multiplanar reconstructions (MPRs), to 5 separate physician readers including 2 radiologists and 3 nonradiologists. These readers graded the visualization of a variety of normal and up to 5 pathologic lesions per patient on a scale of 1 to 5 (5 = best).nnnRESULTSnThe addition of sagittal and coronal multiplanar reconstructions significantly improved the visualization of all normal anatomic structures (mean axial only, 3.8; mean axial plus MPR, 4.1; P < 0.0001). In addition, most pathologic lesions were statistically better visualized with the addition of multiplanar reconstructions (mean axial images only, 3.9; mean axial plus MPR, 4.1; P < 0.0001). All five readers found improved visualization in nearly every category with the addition of the multiplanar reconstructions. However, in only 7% of cases, did a reviewer find new diagnostic information with the addition of MPR images.nnnCONCLUSIONSnStacked multiplanar reconstructions of routine, thick-section abdominal CT has clinical value in both the display of normal anatomic and pathologic lesions. Further studies, however, are required to confirm these findings before it is commonly used.

The effect of helical CT on X-ray attenuation

PURPOSEnConventional CT has been shown to have wide variability in measured CT attenuation, both temporally within the same scanner and between different scanners. Many radiologists have raised the concern that the increased noise and multiple variables associated with helical CT may lead to degradation in resolution, specifically causing errors in CT number values. This study was designed to specifically evaluate the performance of both types of CT scanning in this regard.nnnMETHODnA Picker PQ2000 helical CT scanner was used to scan a phantom containing multiple tissue-equivalent densities, allowing the measurement of CT attenuation of soft tissue, distilled water, cortical bone, medullary bone, air, and fat with a variety of techniques. A Catphan phantom was imaged with a variety of slice thicknesses (2, 4, and 8 mm), phantom positions (isocenter, y = +20 cm), and pitches (1.0, 1.5, 2.0) using both conventional and helical sequences. The entire image set was repeated with two additional annuli placed around the Catphan phantom to simulate the abdomen and the calvarium. The attenuation measurements of the same imaging parameters for helical versus conventional CT were statistically compared.nnnRESULTSnNo statistical differences were found for the CT numbers based on scan type (conventional versus helical) for all sequences and gantry positions tested, including helical CT with pitches > 1.0. Greater CT number variability was found with the extremes of tissue density such as with air and especially cortical bone, but were not statistically significant. The addition of the abdominal and calvarial annuli created a greater variation in CT attenuation values, but again were not statistically significant.nnnCONCLUSIONnThe measurement of X-ray attenuation does not vary significantly with the use of the helical technique.

The optimization of helical thoracic CT

PURPOSEnOur purpose was to determine the optimal helical thoracic CT scanning protocol.nnnMETHODnThree adult Suffolk sheep under general anesthesia were repeatedly scanned by a variety of variable thickness helical and conventional plus thin section high resolution (lung gold standard) CT sequences, reconstructed for mediastinal (standard interpolator and algorithm) and lung parenchymal (extrasharp interpolator, bone algorithm) detail. The images were evaluated in a random order by five separate blinded, experienced imagers utilizing a predetermined grading scale.nnnRESULTSnAt equivalent slice thicknesses, the mediastinal images showed no statistically significant differences between conventional and helical CT using pitches of 1.0, 1.5, and 2.0. However, the 5-mm-thick sections, regardless of technique, performed better than did either the 2- or the 10-mm-thick section images. For the lung interstitium, there was an obvious and marked advantage to reconstructing the lung images separately from the mediastinal images with edge-enhancing algorithms and interpolators. With 1-mm-high mA thin section, high resolution lung CT as the gold standard, 2 mm conventional and helical pitch 1.0, 1.5, and 2.0 images were all graded equivalent. Of the 5 mm images, the helical pitches of 1.0 and 1.5 were graded equivalent to the gold standard. All of the 10 mm lung sections using both conventional and helical CT were graded statistically worse than the gold standard (p < 0.05).nnnCONCLUSIONnThe use of helical CT with a 5 mm beam collimation and a pitch of 1.0 or 1.5 reconstructed twice to maximize both the mediastinal and the lung parenchymal detail provides the optimal way to routinely evaluate the chest.

Utility of low mA 1.5 pitch helical versus conventional high mA abdominal ct

The objective of this study was to evaluate the utility of a low mA 1.5 pitch helical versus conventional high mA conventional technique in abdominal computed tomography (CT). Twenty-five patients who had both a conventional high mA (> 300) and a 1.5 pitch low mA (80-125) helical CT within 3 months were selected for inclusion in the study. Patients were excluded who had a significant change in pathology between the two studies. The other parameters (injection rate, contrast type and volume, and filming window/level) were constant. The studies were randomized and blinded to five separate experienced readers who graded the studies by a variety of normal anatomical structures and pathological criteria. Overview questions also assessed noise, resolution, contrast, and overall quality. The abdominal wall/retroperitoneum and hiatal hernias were statistically better visualized on the conventional high mA studies. However, for all other normal anatomical and pathological sites, there was equivalent or better visualization on the helical versus the conventional CT examinations. The resolution of the helical studies was graded statistically better than the high mA conventional CT scans as was the amount of noise present on the images. While there was some advantage for conventional high mA CT with respect to contrast enhancement and low contrast sensitivity, these differences were not statistically significant. It appears from the data of this study that a low mA technique in evaluating the abdomen may be a useful option in performing routine abdominal CT. The radiation dose savings to the patient is significant and there appears to be little degradation of image quality using a low mA 1.5 helical versus mA conventional CT technique.