This action might not be possible to undo. Are you sure you want to continue?
New Prospective for Non-Invasive Detection,Grading, Size Evaluation, and T umor Location of Prostate Cancer
M. Cortesi,1* E. Fridman,2 A. Volkov,2 S. Sh. Shilstein,1 R. Chechik,1 A. Breskin,1 D. Vartsky,3 G. Raviv,4 and J. Ramon4
Departement of Particle Physics,Weizmann Institute of Science, Rehovot, Israel 2 Institute of Pathology Sheba Medical Center,Tel Hashomer, Israel , 3 NDT Soreq NRC,Yavne, Israel , 4 Department of Urology Sheba Medical Center,Tel Hashomer Israel , ,
BACKGROUND. PSA blood test and other present screening tools fail to provide the required sensitivity and speciﬁcity and, at early stages, lack correlation with tumor grade, volume, and location. Thus alternative approaches are highly desired. We present and assess a novel method for PCa detection, grading, volume evaluation and tumor location, based on non-invasive zinc concentration mapping in the gland by means of a dedicated rectal probe. METHODS. Zinc-concentration values measured in histologically examined tissue fragments from needle biopsy of 598 patients were analyzed. They were used to generate computer simulated zinc-concentration maps, further analyzed with image-processing tools. The tumor detection performances versus Gleason grade were assessed. RESULTS. A signiﬁcant increase of zinc depletion with increasing Gleason pattern (grade) classiﬁcation was established. Tumor detection performance in zinc-concentration maps progressively improves with the cancer’s ﬁrst component score. Reliable information on the location, size and Gleason-grade combination of the lesion can be extracted for clinically relevant volumes. CONCLUSIONS. Zinc depletion in the prostate peripheral zone is the basis for a novel, noninvasive PCa detection, localization, volume evaluation and grading method. Its realization and application as a pre-biopsy and pre-treatment examination, or a follow-up tool, relies on the development of a dedicated transrectal probe. It should have signiﬁcant impact on biopsy effectiveness, point at a possible extraprostatic extension and provide critical data for focal treatment. The information on tumor grade and distribution may have an important impact on disease management. Prostate 70: 1701–1708, 2010. # 2010 Wiley-Liss, Inc. KEY WORDS: prostate cancer; prostatic zinc concentration; patient selection for biopsy; PCa detection; local therapy control
INTRODUCTION Recent publications in the New England Journal of Medicine [1–3] question the effectiveness of PSA screening in reducing death from prostate cancer (PCa). Although these reports from large randomized trials performed in North America and Europe generated somewhat conﬂicting results, they both point at the remarkable deﬁciency of this test. Analysis of the North ß 2010 Wiley-Liss, Inc.
M. Cortesi and E. Fridman contributed equally to this work. *Correspondence to: M. Cortesi, Department of Particle Physics, The Weizmann Institute of Science, 76100 Rehovot, Israel. E-mail: email@example.com Received 18 March 2010; Accepted 6 May 2010 DOI 10.1002/pros.21205 Published online 16 June 2010 in Wiley Online Library (wileyonlinelibrary.com).
 contains a comprehensive report on the concept and tools we developed to extract clinically relevant information from a two-dimensional zincconcentration map. [14. In the present study we rely on the experimental data published in Ref. In the latter. PCa was identiﬁed and classiﬁed according to the Gleason grading scale and the percentage of area occupied by the total (benign þ malignant) glandular tissue (%Glands) in each segment was estimated.e. and none can distinguish between low-grade indolent cancer and high-grade lethal one . Computer-simulation assessment of the zinc-map analysis Ref. 4 þ 5. most patients are referred to needle-biopsy examination.15] contain details on the materials and methods employed in the clinical study at Sheba and Kaplan Medical Centres. The process of information retrieval was based on a series of image processing steps. 4 þ 4. 4 þ 3. The situation calls for the promotion of alternative approaches. within the entire core ($4 mm3) or within four quarters along its length ($1 mm3 segments). DRE. the image was partitioned into N homogeneous clusters classiﬁed by their average gray-levels.17]. Israel. the biopsy-based Epstein criteria was recently shown  to be falsely optimistic in about 24% of the cases. The slides were scanned at each center by a single senior pathologist. The zinc map was transformed into 8-bit gray-scale image. with a calibrated X-Ray Fluorescence (XRF) instrument. In both medical centers the absolute zinc content was measured. vivo zinc-concentration mapping are given in Refs. They were used here to evaluate in some detail the diagnostic potential of the zincmapping approach. We discuss the impact foreseen for the application of such a new PCa detection method as a pre-biopsy and pre-treatment examination. respectively encompassing 326 and 272 patients referred to needle-biopsy examination. As none of the present major screening tools (PSA. and 5 þ 4). with better detection sensitivity and speciﬁcity to provide more efﬁcient selection of patients to biopsy and with possible guidance of the biopsy needles. while a 20% reduction in mortality was observed in the European study. which were tested on simulated maps. carried out during 2004–2007 and brieﬂy reported below. TRUS) is sufﬁciently speciﬁc. including 158 PCa diagnosed cases.e. [14. and evidences were provided to the local nature of the zinc depletion . dimensions and. and the sole diagnosis means. We used Monte Carlo simulation to generate zinc-maps representing a lesion of given dimensions and grade incorporated within a benign tissue at a random location. as an example. and to their transformation into adenocarcinoma cells in particular [6–11]. 3 þ 4. which serves as a secondary screening test. based on prostatic-zinc mapping. Details on a prospective trans-rectal probe for in The Prostate . The cluster (i. encompassed 598 patients referred to needle biopsy examination. its histological Gleason grade.1702 Cortesi et al. its location. It has been repeatedly conﬁrmed [12–15] that zinc concentration in PCa tissue is markedly lower than the concentration in non-cancer prostate tissues. The latest study of this type . most importantly. we further elaborate on their analysis and reﬁne the correlation between the zinc concentration and the prostate carcinoma Gleason grade. American study showed no reduction in PCa mortality attributable to PSA screening. together with a signiﬁcant false-positive rate of $50%. and then subject to image processing of noise reduction followed by cluster segmentation analysis . METHODS Clinical Assessment of zinc Versus PCa Gleason Grade Refs. and in particular the diagnostic merit of zinc-concentration maps measured in vivo over the gland’s peripheral zone. All the tissue cores were then processed according to standard protocols in the Institutes of Pathology at the respective Medical Centers. While the exact causes for the adenocarcinoma of the prostate are not known yet. and assigning them to a histological classiﬁcation or Gleason grade.. with $75% negative rate. We then employ a simulation-based analysis  to assess the diagnostic value of the zinc concentration data. a group of pixels) present in the zinc-map of the prostatic tissue. . it has become clear that zinc is intimately related to the peripheral zone epithelial cells’ functioning in general. the analysis also provided its area. The present article discusses the prospects of a new method for PCa detection. We demonstrate that a twodimensional map of zinc-concentration values contains reliable information on the presence of a cancer lesion. Heavily relying on biopsy-based detection and grade assignment is dangerous not only because of the known limited sensitivity of this examination (up to 83% in repeated examinations) but also because of the limited tissue volume involved. a group of pixels) with the lowest zinc content in the image was identiﬁed and deﬁned as the ‘‘detected’’ lesion. The entire data were sorted according to the diagnosis and the Gleason grade (3 þ 3. The information was extracted by identifying distinct zinc-depleted areas or features (i..
conﬁrms that the various histograms are non-identical (generally at a conﬁdence level P < 0. mean zinc value for each distribution is reported.The distributions are fitted with lognormal function. b: %Gland distributions grouped according to the same criteria. 1. The statistical Kolmogorov–Smirnov (K–S)  test. The histogram of zinc levels measured in non-PCa tissue segments is also included in Figure 1a. including also a benign combination. and their respective area under the curve (AUC) . a:The distribution of zinc concentration measuredin1mm3 prostate-tissue segments. The measured zinc-concentration values in these tissue fragments are plotted (Fig. RESULTS Figure 1 depicts the data from Sheba medical centre. except for the test between Gleason 3 þ 4 and 4 þ 4. Though the Gleason-related zinc-distributions of Figure 1a show a clear systematic trend..  is by no means unique. also presented in the form of ROC curves. The Prostate . and their distributions are statistically non-identical. at 5% conﬁdence level. namely a zinc-value measurement at a single point (e. This plot provided the basis for the grading and tumorsize information of the ‘‘detected’’ lesion. while in cancerous tissue ones the distribution is broad. The values obtained from the analysis of 500 maps of each Gleason-grade and lesion-size combination.g. their corresponding variances are rather large and they are signiﬁcantly overlapping. The data show a clear systematic increase in zinc depletion with increasing aggressiveness of the prostate malignancy. It implies that a single sampling of the distribution. with a sizeable presence of higher values of %Glands. of increasing zinc depletion with increasing Gleason grade. which is inconclusive due to low statistic).New Prospective for Non-Invasive Detection It should be noted that the algorithm presented in Ref. Following the lesion ‘‘detection’’ in comparison to the original map it was possible to decide whether the ‘‘detection’’ is true or false. as for example machine learning mathematics. which include 636 tissue-segments of 1 mm3 from 65 patients with assessed PCa diagnosis and 1703 assigned Gleason grade. grouped according to the diagnosis and to the assigned Gleason grade. By repeating the process 500 times for each lesion dimensions and grades we obtained a complete evaluation of the true and false detection rates. 1a) in correlation with their respective tissue classiﬁcation and Gleason grade combination. and other image processing approaches. were ﬁtted with a Gaussian. $1 mm3 tissue Fig. Figure 1b depicts histograms of the estimated total glandular component (%Gland) for the same groups of tissue segments. In non-cancerous tissue segments the %Gland is sharply concentrated at a mean value of about 15%. the mean and standard deviation for each grade/size combination were represented on a two-dimensional scatter plot. could be employed t identify zincdepleted regions and assess their parameters.001. We have further investigated the correlation between the zinc-content of the ‘‘detected’’ lesion and the area of the ‘‘detected’’ lesion.
the The Prostate predicted detection quality is unprecedented: for example. In this case the ‘‘detected’’ area can be easily corrected to obtain the ‘‘true’’ area. more often selecting solely high-stage prostate cancer patients.9 already for quite small lesions. themap wasgeneratedon thebasis of the experimentaldistributions ofFigure1a. their AUC values vary largely. However. Using the present image processing algorithm. c:Theresultof the segmentationimage-processing. anditrepresents arealisticmap expectedto bemeasuredwith a future trans-rectalprobe. as will be measured with a trans-rectal zinc-imaging probe. Assuming the image represents a 3 Â 3 cm2 prostatic section. 2a). The ROC curves for PSA and PSAD (PSA density) of the patients in our study . For Gleason grade combinations with the primary (more expansive) component >3. shown in Figure 4b. for the clinically relevant tumor volumes. theblack area is thelowest zinc-contentcluster. depending on the study conditions. showing a 30 Â 30 pixels image containing a 5 Â 5 pixels lesion of grade 4 þ 3 (Fig. A simulated zinc map and its analysis is depicted in Figure 2. ROC curves of PSA and its derivatives (PSA density and free/total PSA among them) were published by many authors [21–23]. is not sufﬁcient to provide reliable diagnostic data. the cluster with lowest average zinc content. corresponding to an approximate tumor of 0. should be adequate. We found that when using the present algorithm the ‘‘detected’’-lesion area deviates from the simulated one by an amount depending on the grade. and Figure 5 shows systematics of the area under the ROC curves. Fig. deﬁned as the ‘‘detected’’ lesion. 3. The result after noise reduction is depicted in Figure 2b.125 cc. the ‘‘detected’’ lesion area is under-estimated by $20% for the higher Gleason-grade combinations and overestimated by up to $50% for the lower Gleason-grade combinations. achieved by mapping the zinc content in a prostate section containing a lesion. is shown in black. of 0. a: 30 Â 30 pixel computer-generated zinc-map representing a benign matrixcontaining a 5 Â 5 pixel tumor of Gleason 4 þ 3. for different lesion sizes and Gleason grades. volume) within the lesion. AUC > 0. a repeated sampling of the same distribution. Figure 2c shows the result after segmentation. as depicted in Figure 3. which point out their poor diagnostic power. b:Themap after noisereductionby filtering. the lesion represents a 0. of $0. The detection quality systematically improves with increasing Gleason grade combination and tumor size. the second lowest zinc-content cluster is shown in gray. and much smaller for 4 þ 4 and 4 þ 5 grades. they are usually Fig.25 cm2 malignant tissue.1704 Cortesi et al. . it was modiﬁed to include statistical noise. The zinc value in each pixel of the map was randomly selected from the distributions of Figure 1a for the benign and the malignant components. However. 2.54 and 0. when the lesion extends over many pixels. for example. with the additionofnoise. the gray area is the secondlowest zinc-contentcluster. have signiﬁcantly lower respective AUC values. thus representing a more realistic map. Figure 4 depicts ROC curves (sensitivity vs.6 cc for 4 þ 3. 1-speciﬁcity) resulting from the analysis of 500 maps.62.
Predicted detection performance of the proposed zincmappingmethod. A possible explanation is the composition of the investigated segments. and 0. The detection performance is unprecedented for lesions having firstcomponent Gleason grade >3.5 cc lesions) of various Gleason grades. 1a) conﬁrm a systematic increase of zinc depletion with increasing Gleason grade combination. The data on the %Glands. has similar. 1b). Most important for our hypothesis is the conﬁrmation  that the zinc depletion is locally correlated with the cancer. 4. which is expected from the data of Figure 1a. based on 30 Â 30 pixels Zinc-concentration maps containing lesions of various dimensions and Gleason grades. and that benign tissue component. . lesions with grade-3 ﬁrst component seem to have low speciﬁcity (see Discussion Section) even for large volumes.65. 0. shows that for Gleason 3 þ 4. data from Ref.4.2 and 0.55. The clinical results (Fig. the false-positive rate of zinc-mapping for 0.New Prospective for Non-Invasive Detection 1705 Fig. high zinc content distributions. and the basis for the hypothesis that zinc may be regarded as a speciﬁc natural bio-marker of PCa presence and a predictor of its malignancy and proliferation potential. 5. . respectively for PSA. the zincmapping method performs better than PSA for all lesion-sizes and better than PSAD for tumors larger than 100 mm2 (1 cc). based on 30 Â 30 pixels zincconcentration maps containing 6 Â 6 (a) and 8 Â 8 (b) pixels lesions (respectivelyequivalent to 0. Regrettably. but there is no apparent correlation with the Gleason score.8. free/ total PSA and PSAD. which does not support the use this parameter for diagnostic purpose.1.9). This is a manifestation of the zinc-related metabolic transformation of cellular activity in the epithelial tissue. . and 4 þ 5. representing the AUC versus lesion area. whether taken from a cancerous organ or not. the false-positive rate of PSA is about 0. The Prostate . ROC curvesrepresenting the predicted detection performance of theproposed zinc-mappingmethod. around 0. presented in (Fig. 0. But Figure 5.7. for Gleason 3 þ 4 the zinc-mapping method performs better than PSA for all lesion-sizes and better than PSA-density for tumors larger than100 mm2 (1cc). Figure 4 demonstrates the advantage of the zincbased approach as compared to PSA: To detect 90% of the tumors (sensitivity 0. DISCUSSION In the present article we analyze data form a clinical study described in Ref.The AUC of PSA and PSA-density from Ref. and investigate the diagnostic potential of mapping zinc concentration in the prostate. responsible for the peak at $15% present in all the Fig. PSA and PSA-Density (PSAD) ROC curves are given as well. and 0 for the respective Gleason grades 4 þ 3.5 cc tumor is 0. indeed give an indication of a hyper-proliferation of malignant epithelial glands. which typically contained only a small volume (sometimes a few cells) of cancer and were dominated by the non-cancer component. 4 þ 4.
(The volumes refer to lesion area in the maps. limited discrimination can still be done. CONCLUSIONS The search for new prognostic markers for prostate cancer is currently an animated ﬁeld of interest.html.27. though.prostatecancerfoundation. and is in agreement with their distinct clinical conditions and fate . which in all relevant cases was found to be superior to PSA and PSAD. for grade 4 þ 5 with identiﬁed tumor !0. 6. between highand low-grade cancers. for grade 4 þ 4 with identiﬁed tumor !0. quite similar to PSAD in our population. A discussion of the sources for the shortcoming of these data may be found in Ref. itscoordinates on the scheme of this figure will automatically reveal the grade.9354/. data of different grades are partially overlapping. The grading capability improves systematically with the aggressiveness of the disease. are not satisfactory and it is. The image analysis provides a ‘‘detected’’ tumor-area and its average zinc-content (Fig.05 cc. In the region above the dashed line. A key feature of the method is that the most aggressive cancers are most reliably detected. distributions. see text. for example.histoscanning.13C0/. By plotting these coordinates on the scheme of Figure 6 the lesion classiﬁcation.prostatecancerfoundation. and references . according to the grade. it is assigned with zinc value and area.15 cc and for grade 4 þ 3 with identiﬁed tumor !0. emerging from the zinc-map analysis. it is heading for improved Ultra Sound or MRI based imaging modalities (http:// www. The data segregate in clear loci.4983949/k. . 1a).This feature is the basis for the grading capability of the method: after assigning a zinc value andarea to a‘‘detected’’lesion ofunknowngrade. http://www. for example. and other pattern recognition approaches can be used to reveal the zinc-depleted regions in the map.7).1706 Cortesi et al. Judging from Figure 5.org/site/c.4983951/ k. with suggestions for their possible improvement.6 cc. itIWK2OSG/b. as benign or as a malignant of a certain Gleason grade. With smaller segments under investigation.itIWK2OSG/b. Figure 6 is the basis for the grading capability of the zinc-mapping method: When an unknown lesion is ‘‘detected’’ by the image processing of a zincThe Prostate Fig.6–0. a correlation with the Gleason grade could emerge due to the differences in the tissue topology with increasing score. Zinc content versus area of the‘‘detected’’cancer. Following the grade assignment the ‘‘detected’’ area could be further corrected. The pursuit for new PCa screening and detection means is. The detection of Gleason 3 þ 4 is of relatively poor reliability (AUC ¼ 0. on the other hand. The tools developed for zinc-concentration map’s analysis and lesion ‘‘detection’’ were used for the evaluation of the method in terms of expected sensitivity and speciﬁcity. concentration map. http://www.The data are segregatedin distinctloci. and is revealed from the zinc-map analysis. not distinguished from benign tissue.26. in general.4767791/k. The loci are unambiguously separated from each other in the region below the dashed line. in which case the patient is not at high risk. The fact that the zinc-mapping approach differentiates between 3 þ 4 and 4 þ 3 scores is also very important. heading for Molecular Biology/ Genetic markers in blood or urine screening tests (http://www. even for large lesions.itIWK2OSG/b. unless the tumor is very small. including the benign cases.9) detection can be made for grade 5 þ 4 with practically any tumor size. A unique feature of the new method is the possibility for non-invasive tumor grading. as for example by Figure 3.) Obviously. The figure is based on calculations with 500 maps for each grade/size combination. as indicated in the ﬁgure. This feature originates from the correlation between the degree of zinc depletion and the tumor grade.com/index.28]. and their exact dimensions. based on the data of (Fig. Figure 6 depicts the correlation between these two parameters. Within this region. the tools for image processing employed in this work are not unique. each corresponds to a single Gleason grade combination. on one hand. based on all lesion size/grade combinations studied (500 maps for each combination).prostatecancerfoundation.org/ site/c. The detection of Gleason grade 3 þ 3. we claim that a reliable (AUC ¼ 0. 2). The ‘‘detected’’ lesion area should be further corrected to reveal the true area (see text).org/site/c. emerges.71F2/) [25.
Bangma CH. while reducing the number of low-grade cancer and benign patients referred to this examination. Yokochi LA. Riley TL. Aus G. Cryotherapy and High Intensity Focused Ultrasound (HIFU). Church TR. the zinc-based method has a signiﬁcant diagnostic potential and improving performance with increasing Gleason grade. The partial support of The Israel Cancer Association is gratefully accredited. Cortesi warmly acknowledges the Fellowship of the Lombroso Foundation. 35–40% of cancer patients have low-grade disease qualifying for active surveillance . Cancer of the prostate: Early diagnosis by Zinc and hormone analysis. Maattanen L. Mortality results from a randomized prostate-cancer screening trial. The zinc-based imaging approach combines both directions. and could be used at various stages of the disease management. the new method may improve the sensitivity of the biopsy examination by guiding the needle to the zinc-depleted regions. Zappa M. improving patient’s selection to biopsy due to the high speciﬁcity/sensitivity. Weissfeld JL. Kramer BS. 3. This would increase the rate of true positive biopsies above today’s 25% level .New Prospective for Non-Invasive Detection therein) [29–32]. the introduction of the new method could have a signiﬁcant impact on the entire disease management. location and extension is particularly relevant for the consideration of a focal therapy approach. Erbersdobler A. Chia D. Lattouf J-B. Eur Urol 2008. Screening and prostate-cancer mortality in a randomized European study. Ciatto S. and limited size). It has an important expected impact on the disease handling at various stages of its progress. Heinzer H. N Engl J Med 2009. Validation of the contemporary Epstein criteria for insifniﬁcant prostate cancer in European men. Gelmann EP. Prorok PC. size information. extension and location data provided by the zinc method are critical information items. Berenguer A. except for Spectroscopic MRI. Clapp JD. Bill-Axelson A. Buys SS. indolent lesions that do not call for immediate intervention. At the detection and diagnosis stage the zincmapping probe could be used prior to biopsy examination. The probe is expected to efﬁciently identify patients with high grade cancer and refer them to biopsy. . Andriole GL. 2. Lilja H. Schroder FH. Villers ¨¨ ¨ A.360:e18. At the post-therapy stage.33] that will be based on this approach. According to recently accepted strategy  these approaches are mainly considered in case of unifocal cancer of low risk (low grade. Prostate-speciﬁc antigen levels as a predictor of lethal prostate cancer. Huland H. Izmirlian G. Moss SM. N Engl J Med 2009. Jeldres C. low stage. Andren O. Hugosson J. Kvale PA. Berg CD.17. Roobol MJ. non-invasive follow-up tool.P. the fact that biopsy examination is costly and not without complications. Recker F. Kwiatkowski M. Fouad MN. van der Kwast T. ACKNOWLEDGMENTS The work was partially supported by the Horowitz Foundation Internal Grant Program of the Weizmann Institute and by a research grant from Phyllis & Joseph Gurwin Fund for Scientiﬁc Advancement. At the preoperative stage this information may have a considerable impact on operative decisions. Fall K. Suardi N. 99:526–532. Reuther Professor of Research in the peaceful use of Atomic Energy. 6. Rathmell JM. de Koning HJ. In summary. Breskin is the W. possibly guiding the biopsy examination. Screening for prostate cancer. Crawford ED. Miller AB. which is an expensive examination currently used only in pre-operation phase for staging and could potentially provide grading information. Lee TH. O’Brien B. that $75% of biopsy examinations are negative  and further. Karakiewicz PI. Its potential features include non-invasive detection. such information provided at an early phase of the diagnosis chain could be very important since the majority of PCa cases are low-grade. Furthermore. the data on tumor grade.360:1310– 1319. none of these new imaging means has the potential to provide information on the tumor grade. Adolfsson J. the method could be applied as an effective. Grading the disease and assessing the tumor extent. Haese A. 5. However. Pinsky PF. and grading of lesions of clinically relevant size. current tests suffer from very low sensitivity and speciﬁcity and both current and future tests do not provide any staging or grading information. at the earliest possible phase is of prime importances regarding decisions on further examinations and treatment. Hutterer GC. contract No 5089. among them are Local The Prostate 1707 Brachytherapy. A. Denis LJ. While a blood or urine test is a simple and possibly an economic approach for screening large population. Johansson JE. The trans-rectal probe [14.360:1320– 1328. Br J Cancer 1979. Stitch SR. J Natl Cancer Inst 2007. Auvinen A. Reding DJ. within an imaging modality. Mason MK. McNaughton-Collins MF. Holmberg L. Blijenberg BG. extension and grading information. Adami HO.39:700–704. Grubb RL. REFERENCES 1. Gohagan JK. The imaging approaches have the potential beneﬁt of providing location information. Lujan M. Moreover. Smith PH. localization. Garmo H. Ahyai S. intimately related to the adenocarcinoma onset and progression and strongly correlated to the cancer grade. Walz J. Habib FK. Rebillard X.54:1306–1313. In view of the current over-detection and over-treatment of prostate cancer . Hayes RB. M. It incorporates the natural bio-marker. Graefen M. is expected to provide location. ´ 4. Kantoff PW. Tammela TLJ. ¨ Nelen V. In this case the grade. N Engl J Med 2009.
Nakanouchi T. Djavan B. 11. Eur Urol 1998. Okihara K. Partin AW. 12. Meijer GA. Marks LS. Fridman E. Zaichick VY. Feng P. Urology 2001.40:200–207. Boes DC. Vartsky D. Role of zinc in the pathogenesis and treatment of prostate cancer: Critical issues to resolve. Zelefsky MJ. Murai M. Cannon GW. Breskin A. Costello LC. et al. Franklin RB. Keﬁ A. Loening SA. Volkov A. Tanimoto A. 26. and local extension of prostate cancer. Imai Y. Vartsky D. Chan TY. Zinc causes a shift toward citrate at equilibrium of the m-aconitase reaction of the prostate mitochondria. Kadesky K.70:914–921. Int J Urol 2002. Kojima M. New York: McGraw-Hill. Hara I. Miki T. Ramon J. 31. 29. Kennedy MC.33:261–270. Huszar M. Aslan G. Hamatake M. Heinau M. J Cancer Res Clin Oncol 2007. Lenz A. Analysis of three-dimensional ultrasound Doppler for the detection of prostate cancer. Till RJ. Value of prostate speciﬁc antigen a1-antichymotrypsin complex for the detection of prostate cancer in patients with a PSA level of 4. J Inorg Biochem 2000. Roemer A. Roehrborn CG. Evaluating the cancer detection and grading potential of prostatic-zinc imaging: A simulation study. Chechik R. The value ¨ ¨ ˘ of PSA. Moriel E. Fine SW. Milon B. Nomi M. Focal therapy for localized prostate cancer: A critical appraisal of rational and modalities. Zou J. Esen A. Jiang Z. van der Poel HG. free-to-total PSA ratio and PSA density in the prediction of pathologic stage for clinically localized prostate cancer. Korotinsky S. Mol AJM. Volkov A. Induction of necrosis by zinc in prostate carcinoma cells and identiﬁcation of proteins increased in association with this induction. Ishida R. Chechik R. Fridman E. Shilstein SS. Leman ES. Franklin RB. Horiguchi Y. Yorukoglu K. Invivo determination of prostatic zinc: Phantom feasibility study.54:781–796. Method and system for detecting and grading prostate cancer. van Moorselaar RJ. Kogan G.56:823–827. 28. Kravchick S.57: 1128–1132. Raviv G. Hara S. 21. Mukai M. Endorectal MRI for prediction of tumor site.35:285–296. 30. Shilstein SSh.29:687–694 (and references therein). Prostatic zinc and PSA: An experimental evaluation of their combined diagnostic value. Graybill FA. Int Urol Nephrol 2005. 10. Abenhaim M. J Urol 2007. Walsh PC. Ohigashi T. 1974. 15. Urology 2000. 14. Muge F. 16.78:161–165. antigen (PSA) to prostate volume (PSA density) as a parameter to improve the detection of prostate carcinoma in PSA values in the range of <4 ng/mL. 25. Moriel E. Zaichick SV. Mood AM.1–10. Prostate 1988.133:687–695. Raviv G. Chan DW. Breskin A.37:511–514. Vartsky D. Zinc in the human prostate gland: Normal.178:2260–2267. tumor size. Volkov A. Moskalik A. Eggener SE. 8. Bostwick DG. Gladysh V. Cortesi M. 34. Ramon J. 17.1923:319–323. Alpha-methylacyl-CoA racemase: A multiinstitutional study of a new prostate cancer marker. The ratio of prostate-speciﬁc The Prostate . Oya M.45:218–225. Peled R. Epstein JI. Phys Med Biol 2009. Ben-Dor D. Malnick SD. New experimental markers for early detection of high-risk prostate cancer: Role of cell–cell adhesion and cell migration. Mangold L. Histopathology 2004. 9. Urology 2004. BJU Int 2003. The novel role of Zinc in the regulation of citrate metabolism and its implications in prostate cancer. Ramon J. Introduction to the theory of statistics. Kleinman N. Irer B. Ramos V. Possible use of transrectal power Doppler imaging as an indicator of microvascular density of prostate cancer. Nakashima J. Liang JY. Chechik R. World Intellectual Property Organization (WIPO). 23. Colour Doppler ultrasonography for detecting perineural invasion (PNI) and the value of PNI in predicting ﬁnal pathological stage: A prospective study of men with clinically localized prostate cancer. Trock BJ. Pros Cancer Prost Dis 2004. Miyake H. Partin AW. Urology 2001. Breskin A. Sartor O. Liu Y. Bercovich A. Schnorr D. A simple generalization of the area under the ROC curve to multiple classiﬁcation problems. Raviv G. Takamatsu T. 7. Kuroiwa K. EPCA-2: A highly speciﬁc serum marker for prostate cancer. Hand DJ. Scardino PT. 32. Prognostic signiﬁcance of Gleason score 3þ4 versus Gleason score 4þ3 tumor at radical prostatectomy. Arakawa S. Moriel E.64:101–105. Celebi I. Kravchenko Y. Fridman E. Mach Learn 2001. Sviridova TV. Shilstein SSh. Kogan G.68:994–1006.69:714–720.0 ng/mL: Comparison with PSA-related parameters. Costello LC. Ukimura O. Yokoyama K. 19. Carroll PR. Cancer 2005. Stroebel G. Iguchi K. Map segmentation by colour cube genetic Kmean clustering.170:2258–2262. Marberger M. Wheeler TM. Costello LC. Franklin RB. Kleinman N.45:171–186.92:28–31. Jung K. Ohori M. 35. Rev Urol 2008. Vartsky D. Eur J Biochem 1998. Hricak H. Chechik R. Chechik R.1708 Cortesi et al. Huszar M. 18. Feldman G. July 2009. Lein M. Gleason scores from prostate biopsies obtained with 18-gauge biopsy needles poorly predict Gleason scores of radical prostatectomy specimens. Trachtenberg J. J Phys Med Biol 2004. Urology 2007.8:589–593. International (PCT) patent application. Franklin RB. 33. 22. Cortesi M. number WO 2009/ 083988. Marumo K. Huszar M. Nakagawa K. Rubin MA.49:1–15. Inhibitory effect of zinc on human prostatic carcinoma cell growth.58:573–577. Kojima M. Lect Notes Comput Sci 2000. 3rd edition. Prostate 1999. Breskin A. Prostate 2008. Talanta 2006. Getzenberg RH. Sokoll LJ. Tan M. 27. Vartsky D. Rossignol M.104:993–1003. Prostate cancer speciﬁcity of PCA3 gene testing: Examples from clinical practice. Cortesi M. Prostatic Zn determination for prostate cancer diagnosis. 24. Int Urol Nephrol 1997. Gladysh V. Cytron S. Costello LC. Stephan C. Breskin A. Shilstein S. Liu YY. 13. Klopukh B.7:111–117 (and references therein).10:175– 181. Geldof AA. J Urol 2003. hyperplastic and cancerous. Feng P. Clinical assessment of the cancer diagnostic value of prostatic zinc: A comprehensive needlebiopsy study. Kamidono S.253:766– 770. 20.