DOI: 10.1148/radiol.2422060219
(Radiology 2007;242:637-639.)
© RSNA, 2007
Radiology Research and Rock Soup
F. Graham Sommer, MD
Department of Radiology, Stanford University Medical Center, 300 Pasteur Drive, H-1307, Stanford, CA 94305-5105
e-mail: gsommer{at}stanford.edu
Editor:
In an ancient fable, a traveler is described having wondrous skills, one of which is the preparation of a soup made by using rocks. The townspeople gather for a demonstration by the traveler, who boils water and inserts some carefully chosen rocks. After boiling the rocks for some time, he tastes the preparation and announces that it is very good, but it could use perhaps a little broth, which is provided by an obliging townsman. After a while longer, and another tasting, he says it is almost there, but it could use just a bit of meat. Two more tastings are involved, with the traveler voicing enthusiasm for the quality of the soup but requesting only a bit of meat, then of spices to finish off his creation. Finally, the traveler is satisfied and dispenses samples of the soup to the assembled townspeople, who by now are quite hungry in anticipation. "It is amazing," the townspeople exclaim, "the stranger made such a wonderful soup using rocks!"
I have continually come across what I call the "rock soup argument" in radiology presentations and the radiology literature, and it is one we should guard against if we want to know the truth. The argument is generally made to support the use of an imaging test, the value of which may be undemonstrated or controversial. The general concept often presented is that the value of a test cannot be discerned in isolation but only in the context of all the available clinical information. While it is certainly true that we generally read imaging studies in the presence of all available clinical information, this truth is irrelevant to the design of a study to determine the true incremental value of an imaging test. To determine the independent incremental value of an imaging test, it must be interpreted without the benefit of information known to correlate with the diagnosis one is making with the imaging test in question.
I believe I detect the "rock soup argument" in the article by Dr Wang and colleagues (1), in the February 2006 issue of Radiology, in which the authors describe a study of the incremental value of magnetic resonance (MR) imaging–MR spectroscopic imaging in the staging of prostate cancer. In this retrospective study, MR imaging–MR spectroscopic imaging studies were interpreted in the presence of available clinical data, including prostate-specific antigen (PSA) levels, Gleason tumor grades, and clinical staging; these parameters are all known to correlate with prostate cancer staging as expressed in the innovative Partin tables. While it is certainly true that images obtained at MR imaging–MR spectroscopic imaging may generally be interpreted in the presence of available clinical data, the design of the study by Dr Wang and colleagues is not appropriate for the determination of the incremental value of MR imaging–MR spectroscopic imaging, which is claimed in the article. While it is impossible for us to know the extent to which interpretation of the images was influenced by knowledge of the clinical data known to correlate with tumor stage, it could well have had considerable influence, particularly in difficult cases (which were likely many). It is possible, therefore, that while the benefit of MR to the staging of prostate cancer may be significant, it is also possible that its value may be similar to the contribution of the rocks to the fabled soup. Assessment of the incremental value of MR imaging–MR spectroscopic imaging to prostate cancer staging, if any, will require a study using observers blinded to data known to correlate with prostate cancer stage.
 |
REFERENCE
|
|---|
- Wang L, Hricak H, Kattan MW, Chen HN, Scardino PT, Kuroiwa K. Prediction of organ-confined prostate cancer: incremental value of MR imaging and MR spectroscopic imaging to staging nomograms. Radiology 2006;238:597–603.[CrossRef][Medline]
Response
Peter T. Scardino, MD*,
Hedvig Hricak, MD, PhD
,
Michael W. Kattan, PhD* and
Liang Wang, MD
Departments of Urology* and Radiology, Memorial Sloan-Kettering Cancer Center, 1275 York Avenue, C-278, New York, NY 10021
e-mail: wang6{at}mskcc.org
Dr Sommer has an excellent sense of humor. His tale of "Rock Soup" is amusing. We did state, in the materials and methods section of our article (1) and again in the discussion, that MR readings were prospective clinical interpretations and not retrospective as implied in the letter. The readers had access to available clinical information because we wished to study routine clinical practice. However, at the time of prospective reading, detailed clinical information was not always available, and the consistency with which elements of clinical information were available and used for image interpretation was not recorded. Clinical practice was also reflected in the staging nomogram results, as assignment of clinical stage and performance of endorectal-guided biopsies (which form the basis for histologic diagnosis) were done by a number of different urologists in the clinic.
By using logistic regression analysis and receiver operating characteristic methodology, we evaluated the incremental value of endorectal coil MR imaging and MR spectroscopic imaging to the staging nomograms for predicting organ-confined prostate cancer. If we understand Dr Sommer's letter correctly, he suggests that it was probably the radiologists' awareness of clinical variables, and not their examination of the MR imaging studies, that rendered MR findings of significant incremental value to the Partin staging nomogram. However, while for some diseases clinical information may affect the outcome of image interpretation, clinical findings in prostate cancer are often of limited value. For example, the sensitivity of digital rectal examination for organ-confined disease is only 52% (2), and PSA in most cases does not provide sufficiently accurate staging information because of substantial overlap between PSA levels in different stages (3). Furthermore, Gleason scores correlate closely with stage only at either extreme of the system (scores of 2–4 or 8–10), whereas most patients in our study had a Gleason score of 6 (4).
Prostate cancer staging nomograms were introduced because single variables were found to be inadequate for predicting stage. Staging nomograms are helpful clinical tools that continue to evolve, indicating advances in laboratory and clinical medicine. The Partin staging nomogram, which is based on clinical stage, Gleason score, and serum PSA level, was first published in 1993 and was updated in 1997 and again in 2001 to reflect changes in the presentation and pathologic stage in men newly diagnosed with prostate cancer (5). Other nomograms have been developed to predict stage, individual features of stage (eg, seminal vesicle invasion, extracapsular extension), or recurrence, by using variables such as the extent and location of cancer in systemic biopsy results, the percentage or number of biopsy cores positive for cancer, or the highest biopsy Gleason sum (6–9). However, it has not been established that any of these variables contribute significant incremental value to those used in the Partin nomogram. Thus it cannot be assumed that our radiologists' mere awareness of any of these additional clinical variables at the time of imaging would have rendered MR imaging findings of significant incremental value to the Partin staging nomogram.
In addition, studies in which readers have been blinded to clinical data have shown levels of accuracy for MR imaging similar to those found in our study (10–12). Mullerad et al (13) showed that when MR image readers were blinded to clinical information, MR imaging significantly increased the accuracy of prostate cancer localization by means of digital rectal examination or transrectal ultrasonography guided biopsy (P < .001 for both).
In conclusion, cancer is a heterogeneous disease, and there is no single variable predictive of cancer behavior or indicative of the best cancer treatment for an individual patient. The clinical care of a patient always involves more than the sum of the variables involved, and data from MR is no exception. Our study was designed to inquire about the incremental value of MR imaging and MR spectroscopic imaging in daily clinical practice as a reliable way to inform clinical staging in men with clinically localized prostate cancer. Certainly our results need to be verified by other institutions and ideally in a large multi-institutional study. Such studies, however, should be designed with the knowledge that the quality of the images and the experience and training of the interpreter play an essential role in determining whether this imaging modality can provide value in the care of patients with prostate cancer—a disease notoriously hard to visualize within the heterogeneous prostate.
|
References
|
|---|
- Wang L, Hricak H, Kattan MW, Chen HN, Scardino PT, Kuroiwa K. Prediction of organ-confined prostate cancer: incremental value of MR imaging and MR spectroscopic imaging to staging nomograms. Radiology 2006;238:597–603.[CrossRef][Medline]
- Partin AW, Yoo JK, Carter HB, et al. The use of prostate-specific antigen, clinical stage and Gleason score to predict pathological stage in men with localized prostate cancer. J Urol 1993;150:110–114.[Medline]
- Stamey TA, Yang N, Hay AR, McNeal JE, Freiha FS, Redwine E. Prostate-specific antigen as a serum marker for adenocarcinoma of the prostate. N Engl J Med 1987;317:909–916.[Abstract]
- Partin AW, Carter HB, Chan DW, et al. Prostate specific antigen in the staging of localized prostate cancer: influence of tumor differentiation, tumor volume and benign hyperplasia. J Urol 1990;143:747–752.[Medline]
- Khan MA, Partin AW. Partin tables: past and present. BJU Int 2003;92:7–11.[Medline]
- Gancarczyk KJ, Wu H, McLeod DG, et al. Using the percentage of biopsy cores positive for cancer, pretreatment PSA, and highest biopsy Gleason sum to predict pathologic stage after radical prostatectomy: the Center for Prostate Disease Research nomograms. Urology 2003;61:589–595.[CrossRef][Medline]
- Graefen M, Ohori M, Karakiewicz PI, et al. Assessment of the enhancement in predictive accuracy provided by systematic biopsy in predicting outcome for clinically localized prostate cancer. J Urol 2004;171:200–203.[CrossRef][Medline]
- Koh H, Kattan MW, Scardino PT, et al. A nomogram to predict seminal vesicle invasion by the extent and location of cancer in systematic biopsy results. J Urol 2003;170:1203–1208.[CrossRef][Medline]
- Ohori M, Kattan MW, Koh H, et al. Predicting the presence and side of extracapsular extension: a nomogram for staging prostate cancer. J Urol 2004;171:1844–1849.[CrossRef][Medline]
- Yu KK, Scheidler J, Hricak H, et al. Prostate cancer: prediction of extracapsular extension with endorectal MR imaging and three-dimensional proton MR spectroscopic imaging. Radiology 1999;213:481–488.[Abstract/Free Full Text]
- Sala E, Eberhardt SC, Akin O, et al. Endorectal MR imaging before salvage prostatectomy: tumor localization and staging. Radiology 2006;238:176–183.[Abstract/Free Full Text]
- Engelbrecht MR, Jager GJ, Laheij RJ, Verbeek AL, van Lier HJ, Barentsz JO. Local staging of prostate cancer using magnetic resonance imaging: a meta-analysis. Eur Radiol 2002;12:2294–2302.[Medline]
- Mullerad M, Hricak H, Kuroiwa K, et al. Comparison of endorectal magnetic resonance imaging, guided prostate biopsy and digital rectal examination in the preoperative anatomical localization of prostate cancer. J Urol 2005;174:2158–2163.[CrossRef][Medline]
TOP
REFERENCE
References 
