Adaptation of a 3D prostate cancer atlas for transrectal ultrasound guided target-specific biopsy.

Due to lack of imaging modalities to identify prostate cancer in vivo, current TRUS guided prostate biopsies are taken randomly. Consequently, many important cancers are missed during initial biopsies. The purpose of this study was to determine the potential clinical utility of a high-speed registration algorithm for a 3D prostate cancer atlas. This 3D prostate cancer atlas provides voxel-level likelihood of cancer and optimized biopsy locations on a template space (Zhan et al 2007). The atlas was constructed from 158 expert annotated, 3D reconstructed radical prostatectomy specimens outlined for cancers (Shen et al 2004). For successful clinical implementation, the prostate atlas needs to be registered to each patient's TRUS image with high registration accuracy in a time-efficient manner. This is implemented in a two-step procedure, the segmentation of the prostate gland from a patient's TRUS image followed by the registration of the prostate atlas. We have developed a fast registration algorithm suitable for clinical applications of this prostate cancer atlas. The registration algorithm was implemented on a graphical processing unit (GPU) to meet the critical processing speed requirements for atlas guided biopsy. A color overlay of the atlas superposed on the TRUS image was presented to help pick statistically likely regions known to harbor cancer. We validated our fast registration algorithm using computer simulations of two optimized 7- and 12-core biopsy protocols to maximize the overall detection rate. Using a GPU, patient's TRUS image segmentation and atlas registration took less than 12 s. The prostate cancer atlas guided 7- and 12-core biopsy protocols had cancer detection rates of 84.81% and 89.87% respectively when validated on the same set of data. Whereas the sextant biopsy approach without the utility of 3D cancer atlas detected only 70.5% of the cancers using the same histology data. We estimate 10-20% increase in prostate cancer detection rates when TRUS guided biopsies are assisted by the 3D prostate cancer atlas compared to the current standard of care. The fast registration algorithm we have developed can easily be adapted for clinical applications for the improved diagnosis of prostate cancer.

Three-dimensional reconstruction of pulmonary arteries in plexiform pulmonary hypertension using cell-specific markers. Evidence for a dynamic and heterogeneous process of pulmonary endothelial cell growth.

The plexiform lesions of severe pulmonary hypertension (PH) are complex vascular structures composed primarily of endothelial cells. In this study, we use immunohistochemical markers to identify the various cell layers of pulmonary vessels and to identify different endothelial cell phenotypes in pulmonary arteries affected by severe PH. Our computerized three-dimensional reconstructions of nine vessels in five patients with severe PH demonstrate that plexiform (n = 14) and concentric-obliterative (n = 6) lesions occur distal to branch points of small pulmonary arteries. And, whereas plexiform lesions occur as solitary lesions, concentric-obliterative lesions appear to be only associated with, and proximal to, plexiform structures. The endothelial cells of plexiform lesions express intensely and uniformly the vascular endothelial growth factor (VEGF) receptor KDR and segregate phenotypically into cyclin-kinase inhibitor p27/kip1-negative cells in the central core of the plexiform lesion and p27/kip1-positive cells in peripheral areas adjacent to incipient blood vessel formation. Using immunohistochemistry and three-dimensional reconstruction techniques, we show that plexiform lesions are dynamic vascular structures characterized by at least two endothelial cell phenotypes. Plexiform arteriopathy is not merely an end stage or postthrombotic change--it may represent one stage in an ongoing, angiogenic endothelial cell growth process.

Detection of abnormal E-cadherin expression by simulated prostate biopsy.

PURPOSE: Sampling error is an inherent problem of prostate biopsy. Consequently the determination of whether a given carcinoma is clinically significant based on biopsy results is problematic. We assess the dimensions of sampling error and, thereby, provide insight into the potential value of prognostic markers applied to needle biopsies. MATERIALS AND METHODS: We constructed 3-dimensional computer models of 21 prostatectomy specimens, including outlines of carcinomas, regions of abnormal E-cadherin expression and individual Gleason patterns. The 6 random systematic core biopsy technique and modifications were simulated using a computer algorithm. RESULTS: In 6 of 21 cases the area of abnormal E-cadherin expression and/or high grade carcinoma was not sampled on 6 random systematic core biopsy. The areas missed were either small or inconsistently under sampled regions of the prostate. Modifying the placement of biopsy needles improved the detection of these features. In addition, percent tumor in the needle appeared to be well correlated to percent tumor in the prostate (r = 0.891, r2 = 0.642). CONCLUSIONS: To avoid underestimating the aggressiveness of prostatic carcinoma at least 6 biopsies should be taken from each patient. A more extensive sampling is probably not warranted in all patients but it may prove useful in those in whom extent of disease is unclear or whose general health makes treatment decisions difficult. A reliable estimate of tumor volume in the prostatectomy specimen can be made based on relative amount of tumor in the biopsy specimen on an individual basis.

Computer modeling of prostate biopsy: tumor size and location--not clinical significance--determine cancer detection.

PURPOSE: Sampling error is an inherent problem of prostate biopsy, and the determination of clinical significance based on biopsy results is problematic. We quantify the dimensions of these problems by computer simulation. MATERIALS AND METHODS: We constructed 3-dimensional solid computer models of 59 autopsy prostates containing clinically undetected prostate cancer, and performed simulations of the standard prostate biopsy method. RESULTS: Biopsy simulation detected 19 tumors from the 59 prostates, the majority of which were in the most accessible portion of the prostate, the posterior peripheral zone. Using 0.5 cc or greater tumor volume or less than 0.5 cc and Gleason sum 7 or greater as criteria of significance, the model detected 58% (11 of 19) significant tumors and 20% (8 of 40) insignificant tumors. With 0.25 cc or greater tumor volume or less than 0.25 cc and Gleason sum 7 or greater as criteria 15 of 29 significant (52%) and 4 of 30 insignificant (13%) tumors were detected. Among significant tumors defined by either volume criterion there was a statistical difference between detected and undetected tumors in terms of mean tumor volume and mean ratio of tumor volume-to-prostate volume. Among insignificant tumors defined by either criterion there was no such difference. CONCLUSIONS: As much as 20 to 40% of currently detected prostate cancer may be histologically insignificant, as 4 of 19 cancers were detected when 0.25 cc was used as volume determinant of clinical significance and 8 of 19 were detected when 0.5 cc volume was used. These tumors are detected randomly. On the other hand, perhaps only one-half to three-fourths of clinically significant prostate cancers are being detected, and then only because the volume and anatomic location make them hard to miss.

Morphological analysis and classification of latent prostate cancer using a 3-dimensional computer algorithm: analysis of tumor volume, grade, tumor doubling time and life expectancy.

PURPOSE: We estimate the potential clinical significance of prostate cancers found at autopsy provided the individual had lived to the projected lifespan based on life expectancy tables. MATERIALS AND METHODS: We used 3-dimensional computer models of 59 autopsy prostates that contained clinically undetected carcinoma to determine tumor volumes. Using doubling times of 2, 3, 4 and 6 years, carcinoma volumes at autopsy were extrapolated through patient projected lifespans. The carcinomas were then classified as clinically insignificant or significant according to Mayo Clinic criteria. RESULTS: In 13 patients less than 60 years old, using doubling times of 2, 3, 4 and 6 years, clinically significant tumors were identified in 13 (100%), 10 (77%), 7 (54%) and 7 (54%), respectively. In 46 patients 60 years old or greater significant tumors were identified in 32 (70%), 22 (48%), 21 (46%) and 18 (39%), respectively. A statistical difference (p <0.0001) was found between the mean tumor volume (0.20 +/- 0.10 cc) of 43 organ confined carcinomas and the mean tumor volume (3.26 +/- 3.58 cc) of 16 extracapsular tumors. No capsule perforation was found in tumors with Gleason sums of 4 or less. However, capsule perforation was present in 8 of 31 tumors (25.8%) with Gleason sums of 5 or 6, and 8 of 11 tumors (72.7%) with Gleason scores of 7 or 8. CONCLUSIONS: Prostatic carcinomas that remain clinically insignificant throughout life are likely to have doubling times greater than 4 years. The subset of carcinomas that emerge as clinically significant are likely to have doubling times less than 3 years. Therefore, an accurate method to measure doubling time at diagnosis could, provide an objective indicator to guide clinical management.

A 3-D reconstruction algorithm for interpolation and extrapolation of planar cross sectional data.

A three dimensional (3-D) reconstruction algorithm utilizing both linear interpolation and linear extrapolation was developed for the study of human prostatic cancer. The algorithm was validated by comparing the volumes and shapes of original to reconstructed objects. Synthetic objects of known geometry and wax models with shapes characteristic of prostatic carcinomas were assessed with standard planimetry and by the digital interpolation-extrapolation method. Volume and multifocality measurements obtained by reconstructing excised prostate glands using histologic maps obtained from whole-mount sections were tested. The new algorithm provided greater accuracy in determining tumor volumes than conventional methods. This model provides a basis for mathematical analysis of prostate cancer lesions.