- Email: [email protected]
Correlation of Radiographic Imaging and Histopathology Following Cryoablation and Radio Frequency Ablation for Renal Tumors
Oncology: Adrenal/Renal/Upper Tract/Bladder Correlation of Radiographic Imaging and Histopathology Following Cryoablation and Radio Frequency Ablation for Renal Tumors Christopher J. Weight, Jihad H. Kaouk,* Nicholas J. Hegarty, Erick M. Remer, Charles M. O’Malley, Brian R. Lane, Inderbir S. Gill and Andrew C. Novick From the Glickman Urologic Institute and Department of Radiology (EMR, CMOM), Cleveland Clinic, Cleveland, Ohio
Purpose: Followup after radio frequency ablation and cryotherapy for small renal lesions lacks pathological analysis. The definition of successful tumor ablation has been the absence of contrast enhancement on posttreatment magnetic resonance imaging or computerized tomography. We hypothesized that adding post-ablation kidney biopsy would help confirm treatment success. Materials and Methods: From April 2002 to March 2006 a total of 109 renal lesions in 88 patients were ablated with percutaneous radio frequency ablation and from September 1997 to January 2006 a total of 192 lesions in 176 patients were treated with laparoscopic cryoablation. Patients were followed with radiographic imaging and post-ablation biopsy at 6 months. Results: Radiographic success at 6 months was 85% (62 cases) and 90% (125) for radio frequency ablation and cryoablation, respectively. At 6 months 134 lesions (45%) were biopsied and success in the radio frequency ablation cohort decreased to 64.8% (24 cases), while cryoablation success remained high at 93.8% (91). Six of 13 patients (46.2%) with a 6-month positive biopsy after radio frequency ablation demonstrated no enhancement on posttreatment magnetic resonance imaging or computerized tomography. In patients treated with cryoablation all positive biopsies revealed posttreatment enhancement on imaging just before biopsy. Conclusions: We observed a poor correlation between radiographic imaging and pathological analysis. We recommend post-radio frequency ablation followup biopsy due to the significant risk of residual renal cell cancer without radiographic evidence, although to our knowledge the clinical significance of these viable cells remains to be determined. In contrast, radiographic images of renal lesions treated with cryotherapy appeared to correlate adequately with corresponding histopathological findings in our series. Key Words: kidney; carcinoma, renal cell; cryosurgery; radiography
xtirpative surgery, either radical or partial nephrectomy, has demonstrated long-term oncological control for RCC1 and it remains the standard of care. The success of extirpative nephron sparing surgery demonstrated in open partial nephrectomy series2 has led to attempts to replicate open surgery using laparoscopic techniques.3 Although followup is much shorter, 2 laparoscopic series with 5-year followup showed promising oncological results.4,5 The laparoscopic technique has not been largely adopted among the greater urological community because it remains one of the most technically difficult laparoscopic operations performed by genitourinary surgeons and complication rates may be significant even in experienced hands.6,7 Minimally invasive nephron sparing techniques have been explored to treat localized RCC and decrease the morbidity associated with extirpative surgery. Several techniques are
E
Submitted for publication August 14, 2007. Study received institutional review board approval. * Correspondence: Robotic Urologic Surgery, Glickman Urologic and Kidney Institute, Cleveland Clinic/A100, 9500 Euclid Ave., Cleveland, Ohio 44195 (telephone: 216-444-2976; FAX: 216-4452267; e-mail: [email protected]).
For another article on a related topic see page 1627.
0022-5347/08/1794-1277/0 THE JOURNAL OF UROLOGY® Copyright © 2008 by AMERICAN UROLOGICAL ASSOCIATION
under investigation but currently the most used are cryoablation and RFA. In contradistinction to extirpative surgery, in which the entire pathological specimen can be examined to confirm treatment success, cryoablation and RFA have relied only on radiographic imaging to determine treatment success. Historically preoperative MRI or CT have correlated nicely with subsequent pathological findings in extirpative surgery. These imaging modalities have a high overall accuracy for predicting malignancy with a PPV of 80%, and greater than 90% sensitivity, specificity and NPV.8,9 It is generally believed that if a renal tumor does not enhance, it is not a malignancy and does not need to be treated. Physicians have taken such imaging accuracy from its defined role in the diagnosis of malignant renal lesions and applied it in an alternative and to our knowledge untested role of measuring the adequacy of tumor destruction and recurrence after ablative therapy. The belief has been that a renal lesion no longer enhances after ablation, because the tissues have undergone coagulative necrosis and no longer have a blood supply for delivery of intravenous contrast. This measurement of radiographic success has not been rigorously validated with pathological specimens in the ablative setting. Indeed, we are aware of only 2 small series with 310 and 611 patients, respectively, that have even at-
1277
Vol. 179, 1277-1283, April 2008 Printed in U.S.A. DOI:10.1016/j.juro.2007.11.075
1278
RADIOGRAPHIC IMAGING AND HISTOPATHOLOGY OF ABLATED RENAL MASSES Laparoscopic cryoablation is performed with the patient under general anesthesia. The patient is placed in a modified flank position for a transperitoneal approach or a flank position for a retroperitoneal approach. The kidney is mobilized, the renal mass is identified with intraoperative ultrasound and the probes are inserted into the center of the lesion. We aim for a minimum of 5 mm extension of the ice ball beyond the visible tumor margin. We perform a double freeze-thaw cycle and achieve post-thaw hemostasis using an argon beam coagulator and FloSeal™. Patients undergoing cryoablation were admitted for overnight observation in the hospital. Followup protocol included radiographic imaging on postoperative day 1, at 3, 6 and 12 months, and annually thereafter. Radiographic success was defined as no evidence of central or nodular enhancement after treatment. Routine biopsy of the ablated site was performed immediately after 6-month abdominal imaging. In select cases, including 2 in the RFA group and 2 in the cryoablation group, radiographic followup was concerning for cancer due to persistent enhancement or enlargement of the renal mass. These patients underwent immediate extirpative surgery without biopsy and pathological evaluation revealed RCC in all 4. All preablation and post-ablation biopsies were examined by an experienced genitourinary pathologist with hematoxylin and eosin staining. Table 1 lists the reasons that select patients did not undergo 6-month post-ablation biopsy. Demographic and followup data were collected prospectively in an institutional review board approved database and analyzed retrospectively with JMP IN® statistical software. Differences between the 2 groups were calculated using the Wilcoxon rank sum, chi-square or Fisher exact test, as indicated, with p ⬍0.05 considered statistically significant.
tempted to correlate the radiographic appearance of ablated renal masses with pathological outcomes in humans. Radiographic success remains largely an extrapolation of data accumulated in extirpative surgical series. The definition of radiographic success in the ablative setting has been challenged by data suggesting that radiographic imaging may not always correlate with pathological findings in subsequent nephrectomy specimens.11 We hypothesized that posttreatment percutaneous renal biopsy would provide considerable information on post-ablative treatment success. Therefore, we incorporated a 6-month posttreatment biopsy as part of routine followup in patients undergoing ablative therapies at our institution. METHODS From April 2002 to March 2006 a total of 109 consecutive renal lesions in 88 patients were treated with percutaneous RFA and from September 1997 to January 2006 a total of 192 consecutive renal lesions in 176 patients were treated with laparoscopic cryoablation. Patients scheduled to undergo percutaneous RFA generally underwent biopsy with fine needle aspiration and those scheduled for cryoablation underwent TruCut (Bard® Maxcore™) biopsy intraoperatively. Our techniques for RFA and cryoablation have previously been described.12 RFA is usually performed percutaneously with the patient under local anesthesia and mild sedation. The patient is placed prone in a CT scanner and the renal lesion is localized. After preparing and draping in normal sterile fashion a fine needle aspiration is taken from the renal mass under CT guidance. A 25 cm 7.33Fr StarBurst™ FLEX ablation electrode is placed in the lesion and the tines are advanced under CT guidance. Ablation is performed at 200 W, generating a core temperature of 105C at the tip. This target temperature is maintained for 10 minutes. At the completion of the ablative cycle the generator is turned off and tissue temperature is recorded at 30-second intervals. If any of the 5 thermocouple recordings document a temperature of less than 60C, the cycle is repeated. Multiple overlapping ablations are performed depending on anatomical considerations of the renal mass. At the completion of ablation the probes are retracted and the probe tract is ablated as it is removed. Spiral CT is repeated to evaluate any perinephric hematoma. RFA is generally performed on an outpatient basis.
RESULTS Table 2 lists patient demographics and pretreatment characteristics. Preoperative renal lesion characteristics were similar for RFA and cryoablation with regard to malignant or favor malignant pathology on pretreatment biopsy in 75 vs 135 patients (75% vs 70%, p ⫽ 0.33). Median renal lesion size was 2.5 vs 2.4 cm (p ⫽ 0.65). However, significantly more tumors were centrally located in the RFA group compared to those in the cryoablation series (42 or 39% vs 30 or
TABLE 1. Patient demographics and pretreatment tumor characteristics
No. pts Median age (25%–75% range) No. men (%)/women Median cm radiographic tumor size (25%–75% range) No. central (%) No. kidney status (%): Normal contralat kidney Compromised contralat kidney Solitary kidney Solitary remnant No. ablations No. pretreatment biopsy data No. benign (%) No. indeterminate (%): Favor benign Favor malignant No. malignant (%) Wilcoxon rank sum test Pearson chi-square probability
RFA
Cryoablation
88 68 (57.5–75) 74 (68)/35 (32) 2.5 (1.8–3.8) 42 (39)
176 68 (59–75) 138 (72)/54 (28) 2.4 (1.9–3.8) 30 (16)
25 35 31 18 109 99 9 35 15 20 55
(23) (32) (28) (17) (9) (35) (15) (20) (55)
92 53 46 1 192 192 47 39 29 10 106
p Value 0.9151 (Wilcoxon rank sum test) 0.4665 (Pearson chi-square probability) 0.6450 (Wilcoxon rank sum test) ⬍0.0001 (Pearson chi-square probability) ⬍0.0001 (Pearson chi-square probability)
(48) (28) (24) (0.5) (24) (20) (15) (5) (55)
0.0009 (Pearson chi-square probability)
RADIOGRAPHIC IMAGING AND HISTOPATHOLOGY OF ABLATED RENAL MASSES
1279
TABLE 2. Posttreatment success on radiographic imaging or biopsy at 6 months RFA No. 6-mo radiographic followup (%): Not enhancing Peripheral/linear enhancing Central enhancing No. 6-mo biopsy follow-up (%): Pos for atypical cells No. biopsy not completed (%) No. reason biopsy not completed (%):† Anticoagulation Lost to followup Solitary/remnant/chronic renal insufficiency Recurrence/metastatic disease Benign pretreatment biopsy Death before 6 mos No. radiographic success/total No. with 6-mo followup (%)‡ No. radiographic success with biopsy & 6-mo imaging/total No. (%) No. pathological success (%)§
73 52 10 10 37 13 72
Cryoablation
(67) (72) (14) (14) (34) (35) (66)
139 91 35 14 97 6 95
25 (35) 8 (11) 44 (61) 12 (17) 21 (29) 2 (2.8) 62/73 (85) 29/36 (81) 24 (65)
p Value*
(72) (65) (25) (10) (51) (6.2) (49)
0.2204 0.2830 0.0054 ⬍0.0001 0.0054
3 (3.2) 23 (24) 38 (41) 11 (12) 41 (43) 0 125/139 (90) 86/97 (89) 91 (94)
⬍0.0001 0.0311 0.0069 0.3447 0.0082 0.1844 0.6183 0.2248 ⬍0.0001
* Based on Pearson chi-square probability except Fisher’s exact test used for death before 6 months. † There may have been more than 1 reason. ‡ No central or nodular enhancement. § No malignant cells on 6-month biopsy.
16%, p ⬍0.0001). There were more indeterminate pathological results on pre-ablation biopsy in the RFA group (35 or 35% vs 39 or 20%, p 0.0052). Radiographic Results Posttreatment radiographic followup at 6 months was available in 73 (67%) and 139 patients (72%) who underwent RFA and cryoablation, respectively (table 1). The rate of radiographic success, defined as a lack of central or nodular enhancement on post-contrast CT or subtraction imaging MRI, was 85% (62 cases) for RFA and 90% (125) for cryoablation at 6 months of followup. Pathological Results Pathological examination after renal biopsy or subsequent radical nephrectomy was available in 37 (34%) and 97 cases (51%) for RFA and cryoablation, respectively. Most specimens were obtained from percutaneous kidney biopsies of the ablation site at 6 months in the RFA and cryoablation groups (34 or 92% and 95 or 98%, respectively). The remaining pathological specimens were obtained by radical nephrectomy spurred by obvious radiographic persistence of contrast enhancement and/or growth of the ablated lesion. All 4 of these specimens demonstrated RCC on pathological examination. Post-ablation biopsy was not done in 72 cases (69%) in the RFA group and in 95 (49%) in the cryoablation group (p ⫽ 0.0054). Table 1 lists the reasons for not performing biopsy. They varied significantly between the groups. The rate of pathological success, defined as the lack of malignant/atypical cells on post-ablation biopsy or radical nephrectomy, for RFA and cryoablation was 65% and 94% (24 and 91 patients, respectively, p ⫽ ⬍0.0001). Eight of the 10 cases of positive biopsies in the RFA cohort were retreated with percutaneous RFA. Six patients demonstrated no evidence of enhancement on subsequent followup imaging. The remaining 2 re-ablated lesions showed a persistent thin rim of enhancement and they are being followed. The other 2 positive biopsy lesions are being followed with serial imaging, which has revealed stable central enhancing lesions. Two of the 4 patients with positive biopsies in the cryoablation cohort underwent percutaneous RFA, 1
underwent repeat laparoscopic cryoablation and 1 underwent laparoscopic radical nephrectomy, which demonstrated RCC clear cell variant grade 3. In the RFA group 12 of the 13 patients with positive pathological results after ablation had a malignant (8 or 62%) or favor malignant (4 or 31%) diagnosis on pretreatment pathological assessment. However, 1 patient showed benign pathological findings on pretreatment biopsy and malignant pathological findings on followup biopsy. In the cryoablation group only 3 of 6 patients (50%) with pathologically proven RCC after ablation had a malignant (1 or 33%) or favor malignant (2 or 67%) diagnosis on pretreatment pathological evaluation. Correlation of Pathological and Radiographic Results Table 3 shows correlated radiographic and pathological data. For correlative purposes all patients except 1 underwent radiographic imaging with contrast material just before biopsy, including 36 for RFA and 97 for cryoablation. When stratified according to imaging, 6 patients (24%) without any evidence of post-ablation enhancement on 6-month biopsy were found to have viable renal cancer cells in the
TABLE 3. Contrast radiographic and pathological data No. RFA (%)
No. Cryoablation (%)
Overall No enhancement:
36 25 (69.4)
97 60 (61.9)
Neg biopsy Pos biopsy Pos peripheral enhancement: Neg biopsy
19 (76) 6 (24) 4 (11.1)
60 (100) 0 26 (26.8)
(50)
24 (92.3)
Pos biopsy Pos central enhancement: Neg biopsy
2 (50) 7 (19.4)
2 (7.7) 11 (11.3)
Pos biopsy
2
2
(29)
7
(64)
5
(71)
4
(36)
p Value 0.4181 (Pearson chi-square probability) 0.0004 (Fisher’s exact test) 0.0544 (Pearson chi-square probability) 0.0205 (Pearson chi-square probability) 0.2248 (Pearson chi-square probability) 0.1469 (Pearson chi-square probability)
1280
RADIOGRAPHIC IMAGING AND HISTOPATHOLOGY OF ABLATED RENAL MASSES
RFA group. This finding was not observed in the cryoablation group since all 60 patients (100%) without evidence of post-ablation enhancement had negative biopsies. In those with peripheral linear enhancement 2 biopsies (50%) were positive in the RFA cohort compared to only 2 (7.7%) in the cryoablation group (p ⫽ 0.0205). Centrally enhancing tumors yielded positive biopsies in 71% (5 cases) with RFA treatment but only in 36% (4) in the cryoablation group (p ⫽ 0.1469). Six of all positive post-ablation biopsies (46%) in the RFA group showed no enhancement on immediate pre-biopsy imaging, whereas all 6 positive biopsies (100%) in the post-cryoablation cohort came from tumors demonstrating some degree of enhancement. The figure shows an example of a lesion showing no enhancement at 6 months with a positive biopsy for renal cell cancer. There were 2 central enhancing lesions at 6-month followup with a negative post-ablation biopsy in the RFA group (table 3). Enhancement disappeared in 1 patient at 1-year followup and the other is being watched according to patient preference. There were 7 central enhancing lesions at 6-month followup with a negative post-ablation biopsy in the cryoablation group. Three of these patients subsequently demonstrated RCC on repeat biopsy (2) or radical nephrectomy (1). Three other patients are being watched according to patient preference and 1 was lost to followup. There were 2 peripheral enhancing lesions at 6-month followup and a negative post-ablation biopsy in the RFA group. One lesion no longer enhances and the other continues to show stable peripheral enhancement. In the cryoablation group there were 24 lesions that had thin peripheral enhancement at 6-month followup. Of these lesions 15 no longer enhanced on followup imaging, 5 were lost to followup and 2 are being watched with stable peripheral enhancement. Two patients showed radiographic recurrence, defined as a nodular central enhancing lesion. The sensitivity of central nodular enhancement on 6-month followup CT or MRI to predict a positive biopsy in our post-cryoablation cohort was 77.8% with 95.1% specificity, 63.4% PPV and 97.7% NPV. The sensitivity of central nodular enhancement on 6-month followup CT or MRI to detect a positive biopsy in our post-RFA cohort was only 38.4% with 91.3% specificity, 71.4% PPV and 72.4% NPV. DISCUSSION To our knowledge our results are unique because they represent the largest experience available for the correlation between radiographic findings and pathology results after
Example of enhancing renal tumor (A) ablated with RFA and MRI at 6 months shows no enhancement (B). Ablation bed was then biopsied and showed viable renal cell carcinoma.
ablative procedures. Traditionally ablative techniques have been considered successful if followup imaging demonstrated no enhancement or enlargement of the tumor ablation site. MRI has been cited as 99% sensitive for detecting malignant renal lesions with a PPV of 89% and NPV of 91% in the preoperative setting.9 There have been minimal data on the efficacy of MRI or CT to detect tumor destruction in post-ablative cases. Many groups have unquestioningly extrapolated the excellent diagnostic accuracy of CT and MRI to be used as the measurement of success in a post-ablative case. Rendon et al suggested caution when interpreting postablation radiographic imaging when they found that 2 of 6 patients had renal cell cancer at extirpative surgery 1 week later.11 However, this study was criticized for using dry RFA, which creates skip lesions,13 and they did not use NADH staining, which is better than hematoxylin and eosin staining for distinguishing actual viable RCC cells from ablated RCC cells in the early post-ablative setting.14,15 In our study we used the newer StarBurst FLEX RFA probe and performed followup biopsies 6 months after ablation, allowing maturation of the ablation effect and, thus, avoiding the discrepancy between hematoxylin and eosin, and NADH viability stains. Our study addressed each of these criticisms and it appears that caution is still warranted, particularly in the radiographic followup of patients treated with RFA. In the setting of the post-cryoablation patient radiographic and pathological findings seem to correlate adequately. The high NPV demonstrates that we already suspected recurrence by radiographic information in most of these patients and biopsy added little to our treatment algorithm. In contradistinction, the correlation of radiographic and pathological findings in the setting of the post-RFA patient is poor. Our radiographic success rate of 86.1% in the percutaneous RFA cohort is similar to that in other published series of comparable size and central location.16 –20 What is particularly concerning is that almost a quarter of our radiographic successes (6) actually showed a positive biopsy at 6 months. Although our methods should be reproduced at other centers to determine whether this effect and the rate at which it occurs is consistent, there may be a substantial population of patients who are cured by radiographic criteria who have viable RCC cells persisting in the kidney. However, to our knowledge the clinical significance of these viable cells is unknown at this time and longer followup is needed to determine whether these patients are at higher risk for local or systemic progression. Although biopsy added a significant amount of information it, too, has shortcomings. We observed a total of 4 cases in the 2 groups of patients in which initial pretreatment biopsy demonstrated benign pathological results and 6-month posttreatment biopsy revealed malignancy. Renal biopsy alone continues to lack a sufficient NPV to make sound oncological treatment decisions. Our study is limited by the fact that it is retrospective and only about half of the patients underwent posttreatment biopsy. It is also limited by the lack of viability staining of the pathological specimens. As mentioned, analysis of postRFA pathological assessment can be misleading using hematoxylin and eosin staining only. It appears that this difficultly may largely decrease as time from ablation increases. Martin et al found that most patients with positive hematoxylin and eosin staining at 6 months also had positive viability
RADIOGRAPHIC IMAGING AND HISTOPATHOLOGY OF ABLATED RENAL MASSES stains in a hepatocellular carcinoma model.21 Further research addressing these issues is imperative.
11.
CONCLUSIONS
12.
We observed a poor correlation between post-RFA imaging and post-RFA biopsy results at 6 months. To our knowledge the significance of these viable cells on biopsy remains to be determined. We propose that biopsy of the ablated site should be added to RFA followup protocols regardless of pretreatment biopsy results, or the presence or absence of enhancement. In contrast, lesions treated with cryotherapy seem to correlate adequately with radiographic findings and biopsy may not add substantial information.
Abbreviations and Acronyms CT ⫽ computerized tomography MRI ⫽ magnetic resonance imaging NADH ⫽ nicotinamide adenine dinucleotide (reduced form) NPV ⫽ negative predictive value PPV ⫽ positive predictive value RCC ⫽ renal cell carcinoma RFA ⫽ radio frequency ablation
13.
14.
15.
16.
17.
18.
REFERENCES 1.
2.
3.
4. 5.
6.
7.
8.
9.
10.
Frank I, Blute ML, Leibovich BC, Cheville JC, Lohse CM and Zincke H: Independent validation of the 2002 American Joint Committee on cancer primary tumor classification for renal cell carcinoma using a large, single institution cohort. J Urol 2005; 173: 1889. Fergany AF, Hafez KS and Novick AC: Long-term results of nephron sparing surgery for localized renal cell carcinoma: 10-year followup. J Urol 2000; 163: 442. Gill IS, Desai MM, Kaouk JH, Meraney AM, Murphy DP, Sung GT et al: Laparoscopic partial nephrectomy for renal tumor: duplicating open surgical techniques. J Urol 2002; 167: 469. Lane BR and Gill IS: 5-Year outcomes of laparoscopic partial nephrectomy. J Urol 2007; 177: 70. Permpongkosol S, Bagga HS, Romero FR, Sroka M, Jarrett TW and Kavoussi LR: Laparoscopic versus open partial nephrectomy for the treatment of pathological T1N0M0 renal cell carcinoma: a 5-year survival rate. J Urol 2006; 176: 1984. Ramani AP, Desai MM, Steinberg AP, Ng CS, Abreu SC, Kaouk JH et al: Complications of laparoscopic partial nephrectomy in 200 cases. J Urol 2005; 173: 42. Weld KJ, Venkatesh R, Huang J and Landman J: Evolution of surgical technique and patient outcomes for laparoscopic partial nephrectomy. Urology 2006; 67: 502. Hecht EM, Israel GM, Krinsky GA, Hahn WY, Kim DC, Belitskaya-Levy I et al: Renal masses: quantitative analysis of enhancement with signal intensity measurements versus qualitative analysis of enhancement with image subtraction for diagnosing malignancy at MR imaging. Radiology 2004; 232: 373. Hallscheidt PJ, Bock M, Riedasch G, Zuna I, Schoenberg SO, Autschbach F et al: Diagnostic accuracy of staging renal cell carcinomas using multidetector-row computed tomography and magnetic resonance imaging: a prospective study with histopathologic correlation. J Comput Assist Tomogr 2004; 28: 333. Park S, Strup SE, Saboorian H and Cadeddu JA: No evidence of disease after radiofrequency ablation in delayed nephrectomy specimens. Urology 2006; 68: 964.
19.
20.
21.
1281
Rendon RA, Kachura JR, Sweet JM, Gertner MR, Sherar MD, Robinette M et al: The uncertainty of radio frequency treatment of renal cell carcinoma: findings at immediate and delayed nephrectomy. J Urol 2002; 167: 1587. Hegarty NJ, Gill IS, Desai MM, Remer EM, O’Malley CM and Kaouk JH: Probe-ablative nephron-sparing surgery: cryoablation versus radiofrequency ablation. Urology 2006; 68: 7. Rehman J, Landman J and Sundaram CP: Re: the uncertainty of radio frequency treatment of renal cell carcinoma: findings at immediate and delayed nephrectomy. J Urol 2002; 168: 2128. Ogan K and Cadeddu JA: Re: the uncertainty of radio frequency treatment of renal cell carcinoma: findings at immediate and delayed nephrectomy. J Urol 2002; 168: 2129. Crowley JD, Shelton J, Iverson AJ, Burton MP, Dalrymple NC and Bishoff JT: Laparoscopic and computed tomographyguided percutaneous radiofrequency ablation of renal tissue: acute and chronic effects in an animal model. Urology 2001; 57: 976. Gervais DA, McGovern FJ, Arellano RS, McDougal WS and Mueller PR: Renal cell carcinoma: clinical experience and technical success with radio-frequency ablation of 42 tumors. Radiology 2003; 226: 417. Farrell MA, Charboneau WJ, DiMarco DS, Chow GK, Zincke H, Callstrom MR et al: Imaging-guided radiofrequency ablation of solid renal tumors. AJR Am J Roentgenol 2003; 180: 1509. Mayo-Smith WW, Dupuy DE, Parikh PM, Pezzullo JA and Cronan JJ: Imaging-guided percutaneous radiofrequency ablation of solid renal masses: techniques and outcomes of 38 treatment sessions in 32 consecutive patients. AJR Am J Roentgenol 2003; 180: 1503. Su LM, Jarrett TW, Chan DY, Kavoussi LR and Solomon SB: Percutaneous computed tomography-guided radiofrequency ablation of renal masses in high surgical risk patients: preliminary results. Urology 2003; 61: 26. Zagoria RJ, Hawkins AD, Clark PE, Hall MC, Matlaga BR, Dyer RB et al: Percutaneous CT-guided radiofrequency ablation of renal neoplasms: factors influencing success. AJR Am J Roentgenol 2004; 183: 201. Martin AP, Goldstein RM, Dempster J, Netto GJ, Katabi N, Derrick HC et al: Radiofrequency thermal ablation of hepatocellular carcinoma before liver transplantation–a clinical and histological examination. Clin Transplant 2006; 20: 695.
EDITORIAL COMMENT These authors evaluated the reliability of cross-sectional imaging to follow post-ablation renal tumors. To this end I believe that MRI may be less reliable than CT for assessing the adequacy of complete ablation of central tumors after RFA. However, I believe that this is a limitation of the imaging technology and not of the ablation technology. Most clinicians would agree that MRI does not distinguish tumor margins or enhancement as well as CT for most renal tumors. Unfortunately articles such as this may incorrectly provide support for those who advocate the superiority of cryoablation over RFA. As such, I believe that it important to address several methodological issues before any comparison is made. While the authors compared percutaneous RFA to laparoscopic cryoablation, there appears to have been a significant selection bias against RFA. There is no doubt that incomplete ablations are to be expected at a higher frequency in a percutaneous treatment group that includes twice as many central tumors, half as many with a normal contralateral kidney and 17 times more solitary renal remnants. It seems that RFA may have been offered to patients who were not candidates for laparoscopic cryoablation. The treating physician may have been
1282
RADIOGRAPHIC IMAGING AND HISTOPATHOLOGY OF ABLATED RENAL MASSES
more judicious with needle placement and ablation size in such a percutaneous RFA group. Any imaging technology to assess residual disease in central ablations and previously operated kidneys is likely to be more difficult to interpret. Furthermore, since the percutaneous approach is easily repeatable in such high risk groups, incomplete treatment may be more frequent than with the direct vision laparoscopic approach, which cannot be repeated, and it may result in more aggressive laparoscopic ablation. Therefore, the suggestion of superiority for cryoablation based on this study design could well be unfounded. Similarly the inference that CT or MRI can accurately interpret the cryoablation of central tumors or renal remnants may not be possible due to the paucity of such patients in this cohort. As suggested by the authors, the finding of residual viable cells identified on needle biopsy 6 months after RFA has unclear clinical significance in light of good intermediate term clinical data, varying ablation techniques and longer term pathological data collected by others.1-3 Several groups have now reported RFA results at 2, 3 and 5 years of followup, of which none approach the 35% persistent disease, recurrence or metastatic progression rate reported in this study. This discrepancy may be better explained by factors such as tumor selection (as discussed previously), ablation technique (1 RFA cycle instead of 2 routine ablations) and biopsy interpretation. The issue of biopsy interpretation is most important, concerning and difficult to understand. The authors suggest that the timing of biopsy at 6 months allows maturation of the ablation effect and avoids the discrepancy between hematoxylin and eosin, and NADH viability assessments debated in the literature. How do they know this? How do they know that they saw viable cells in the 6 of 25 cases with no enhancement but a positive biopsy? RFA treated lesions are characterized by protein denaturation and crosslinking, which stabilize the tumor architecture (heat fixation, which is similar to frozen section), resulting in a chronic granulomatous response (there is no neutrophil or macrophage infiltration, as with cryoablation) and slow reabsorption, if any. Thus, interpreting the presence of preserved cells in hematoxylin and eosin stained tissue as viable tumor may be incorrect. Instead, they may represent fixed dead cells or granulomatous inflammatory cells, as reported by others (reference 10 in article). In a similar study the University of Michigan group reported that biopsies 2 months after RFA demonstrated residual cancer on hematoxylin and eosin staining in some RFA tumors but NADH staining suggested no viable cells. We have biopsied 20 post-RFA tumors that showed no enhancement on CT, representing a population size similar to that in this study of 25 cases of negative CT or MRI, more than 1 year after RFA and in each there were no visible malignant or atypical cells. Thus, it may take a while for post-RFA heat fixated cells to degenerate, which would account for the atypical cells. Although no one would argue that residual enhancement may indicate cancer, the conclusion of a significant risk of residual renal cell cancer after RFA despite nonenhancement on cross-sectional imaging is only based on 25 cases, of which 6 (24%) had atypical cells. Interestingly in the cases of positive enhancement a similar percent had negative biopsies. Do these patients have residual cancer? In light of the pathophysiological response to thermal ablation, the biopsy experience described and the difficulty with assessing true cellular viability incomplete ablation in 24% of cases has not been confirmed with nonenhancing CT or MRI
and, therefore, convincing evidence for recommending routine needle biopsy after RFA is not provided. The authors report that biopsy after RFA is unreliable. It remains that surveillance with frequent cross-sectional imaging, preferably with CT, is the standard method for following patients after ablation therapy. Needle biopsy should be used in cases of tumor bed enhancement on cross-sectional imaging. With proper patient selection (neither central tumors nor renal remnants) cryoablation and RFA can be expected to have good intermediate term disease control. Jeffrey A. Cadeddu Urology and Radiology Clinical Center for Minimally Invasive Urologic Cancer Treatment University of Texas Southwestern Medical Center at Dallas Dallas, Texas 1.
McDougal WS, Gervais DA, McGovern FJ and Mueller PR: Long-term followup of patients with renal cell carcinoma treated with radio frequency ablation with curative intent. J Urol 2005; 174: 61. 2. Park S, Anderson JK, Matsumoto ED, Lotan Y, Josephs S and Cadeddu JA: Radiofrequency ablation of renal tumors: intermediate-term results. J Endourol 2006; 20: 569. 3. Stern J, Svatek R, Park S, Hermann M, Lotan Y, Sagalowsky AI et al: Intermediate comparison of partial nephrectomy and radiofrequency ablation for clinical T1a renal tumors. BJU Int 2007; 100: 287.
REPLY BY AUTHORS We agree that limitations of imaging do not translate into limitations of ablation, and this has never been our claim. However, when evaluating a new technology, it is prudent to understand the end points used to determine success, especially for patients with potentially lethal disease. Our study raises serious doubts about the ability to determine successful RFA with imaging alone. We believe that these end points deserve research rather than assumption and postulation. Interestingly the comment provides several potential explanations as to why such discrepancies between pathological and radiographic outcomes. We made no attempt to compare the efficacy of RFA and cryoablation but focused only on radiographically successful RFA and cryoablation, and compared how imaging after each ablative modality correlated with post-ablation biopsy. Our findings suggest that post-RFA imaging does not correlate well with post-RFA biopsy. The 3 articles cited in the comment are almost entirely based on radiographic imaging, with little post-RFA pathological data. The first report had no pathological followup data (reference 1 in comment), and the other 2 reports include pathological followup data on 2 cases (references 2 and 3 in comment). We commend Cadeddu for implementing post-ablation biopsy which is exactly the data we need to be able to make informed decisions about ablative techniques. The ability to interpret H&E staining after RFA is a complicated subject. A 6-month biopsy appears to be adequate because there are data in the hepatocellular RFA literature showing that the majority of patients with positive H&E staining at 6 months will also have positive viability stains (reference 21 in article) and these positive biopsies clearly looked different from negative sites. If this was just a fixation effect,
RADIOGRAPHIC IMAGING AND HISTOPATHOLOGY OF ABLATED RENAL MASSES we would expect to see nearly all positive biopsies. Our biopsy findings were not “atypical cells with inflammatory response,” but rather textbook images of RCC. So far, only the RFA literature shows at least debatable inconsistency between H&E and NADH staining, ghost cells, and discrepancy between radiographic and pathological definitions of success. For researchers and clinicians these findings should at least be flagged and further investigated, rather than finding good scenario conclusions up front and ignoring a potential problem in the oncologic adequacy of a given technique.
1283
We agree that cryoablation and RFA for selected small and superficial renal lesions are expected to give good results. However partial nephrectomy remains the gold standard for such tumors, and we disagree that “proper patient selection” excludes central tumors and renal remnants. In fact, this subgroup of patients may benefit the most from thermal ablation when no other surgical management is feasible. Obtaining pathology data after RFA seems even more relevant given the confusion regarding interpretation of imaging as well as the biopsies. We owe our patients the diligence to sort out this issue.
Purpose: Followup after radio frequency ablation and cryotherapy for small renal lesions lacks pathological analysis. The definition of successful tumor ablation has been the absence of contrast enhancement on posttreatment magnetic resonance imaging or computerized tomography. We hypothesized that adding post-ablation kidney biopsy would help confirm treatment success. Materials and Methods: From April 2002 to March 2006 a total of 109 renal lesions in 88 patients were ablated with percutaneous radio frequency ablation and from September 1997 to January 2006 a total of 192 lesions in 176 patients were treated with laparoscopic cryoablation. Patients were followed with radiographic imaging and post-ablation biopsy at 6 months. Results: Radiographic success at 6 months was 85% (62 cases) and 90% (125) for radio frequency ablation and cryoablation, respectively. At 6 months 134 lesions (45%) were biopsied and success in the radio frequency ablation cohort decreased to 64.8% (24 cases), while cryoablation success remained high at 93.8% (91). Six of 13 patients (46.2%) with a 6-month positive biopsy after radio frequency ablation demonstrated no enhancement on posttreatment magnetic resonance imaging or computerized tomography. In patients treated with cryoablation all positive biopsies revealed posttreatment enhancement on imaging just before biopsy. Conclusions: We observed a poor correlation between radiographic imaging and pathological analysis. We recommend post-radio frequency ablation followup biopsy due to the significant risk of residual renal cell cancer without radiographic evidence, although to our knowledge the clinical significance of these viable cells remains to be determined. In contrast, radiographic images of renal lesions treated with cryotherapy appeared to correlate adequately with corresponding histopathological findings in our series. Key Words: kidney; carcinoma, renal cell; cryosurgery; radiography
xtirpative surgery, either radical or partial nephrectomy, has demonstrated long-term oncological control for RCC1 and it remains the standard of care. The success of extirpative nephron sparing surgery demonstrated in open partial nephrectomy series2 has led to attempts to replicate open surgery using laparoscopic techniques.3 Although followup is much shorter, 2 laparoscopic series with 5-year followup showed promising oncological results.4,5 The laparoscopic technique has not been largely adopted among the greater urological community because it remains one of the most technically difficult laparoscopic operations performed by genitourinary surgeons and complication rates may be significant even in experienced hands.6,7 Minimally invasive nephron sparing techniques have been explored to treat localized RCC and decrease the morbidity associated with extirpative surgery. Several techniques are
E
Submitted for publication August 14, 2007. Study received institutional review board approval. * Correspondence: Robotic Urologic Surgery, Glickman Urologic and Kidney Institute, Cleveland Clinic/A100, 9500 Euclid Ave., Cleveland, Ohio 44195 (telephone: 216-444-2976; FAX: 216-4452267; e-mail: [email protected]).
For another article on a related topic see page 1627.
0022-5347/08/1794-1277/0 THE JOURNAL OF UROLOGY® Copyright © 2008 by AMERICAN UROLOGICAL ASSOCIATION
under investigation but currently the most used are cryoablation and RFA. In contradistinction to extirpative surgery, in which the entire pathological specimen can be examined to confirm treatment success, cryoablation and RFA have relied only on radiographic imaging to determine treatment success. Historically preoperative MRI or CT have correlated nicely with subsequent pathological findings in extirpative surgery. These imaging modalities have a high overall accuracy for predicting malignancy with a PPV of 80%, and greater than 90% sensitivity, specificity and NPV.8,9 It is generally believed that if a renal tumor does not enhance, it is not a malignancy and does not need to be treated. Physicians have taken such imaging accuracy from its defined role in the diagnosis of malignant renal lesions and applied it in an alternative and to our knowledge untested role of measuring the adequacy of tumor destruction and recurrence after ablative therapy. The belief has been that a renal lesion no longer enhances after ablation, because the tissues have undergone coagulative necrosis and no longer have a blood supply for delivery of intravenous contrast. This measurement of radiographic success has not been rigorously validated with pathological specimens in the ablative setting. Indeed, we are aware of only 2 small series with 310 and 611 patients, respectively, that have even at-
1277
Vol. 179, 1277-1283, April 2008 Printed in U.S.A. DOI:10.1016/j.juro.2007.11.075
1278
RADIOGRAPHIC IMAGING AND HISTOPATHOLOGY OF ABLATED RENAL MASSES Laparoscopic cryoablation is performed with the patient under general anesthesia. The patient is placed in a modified flank position for a transperitoneal approach or a flank position for a retroperitoneal approach. The kidney is mobilized, the renal mass is identified with intraoperative ultrasound and the probes are inserted into the center of the lesion. We aim for a minimum of 5 mm extension of the ice ball beyond the visible tumor margin. We perform a double freeze-thaw cycle and achieve post-thaw hemostasis using an argon beam coagulator and FloSeal™. Patients undergoing cryoablation were admitted for overnight observation in the hospital. Followup protocol included radiographic imaging on postoperative day 1, at 3, 6 and 12 months, and annually thereafter. Radiographic success was defined as no evidence of central or nodular enhancement after treatment. Routine biopsy of the ablated site was performed immediately after 6-month abdominal imaging. In select cases, including 2 in the RFA group and 2 in the cryoablation group, radiographic followup was concerning for cancer due to persistent enhancement or enlargement of the renal mass. These patients underwent immediate extirpative surgery without biopsy and pathological evaluation revealed RCC in all 4. All preablation and post-ablation biopsies were examined by an experienced genitourinary pathologist with hematoxylin and eosin staining. Table 1 lists the reasons that select patients did not undergo 6-month post-ablation biopsy. Demographic and followup data were collected prospectively in an institutional review board approved database and analyzed retrospectively with JMP IN® statistical software. Differences between the 2 groups were calculated using the Wilcoxon rank sum, chi-square or Fisher exact test, as indicated, with p ⬍0.05 considered statistically significant.
tempted to correlate the radiographic appearance of ablated renal masses with pathological outcomes in humans. Radiographic success remains largely an extrapolation of data accumulated in extirpative surgical series. The definition of radiographic success in the ablative setting has been challenged by data suggesting that radiographic imaging may not always correlate with pathological findings in subsequent nephrectomy specimens.11 We hypothesized that posttreatment percutaneous renal biopsy would provide considerable information on post-ablative treatment success. Therefore, we incorporated a 6-month posttreatment biopsy as part of routine followup in patients undergoing ablative therapies at our institution. METHODS From April 2002 to March 2006 a total of 109 consecutive renal lesions in 88 patients were treated with percutaneous RFA and from September 1997 to January 2006 a total of 192 consecutive renal lesions in 176 patients were treated with laparoscopic cryoablation. Patients scheduled to undergo percutaneous RFA generally underwent biopsy with fine needle aspiration and those scheduled for cryoablation underwent TruCut (Bard® Maxcore™) biopsy intraoperatively. Our techniques for RFA and cryoablation have previously been described.12 RFA is usually performed percutaneously with the patient under local anesthesia and mild sedation. The patient is placed prone in a CT scanner and the renal lesion is localized. After preparing and draping in normal sterile fashion a fine needle aspiration is taken from the renal mass under CT guidance. A 25 cm 7.33Fr StarBurst™ FLEX ablation electrode is placed in the lesion and the tines are advanced under CT guidance. Ablation is performed at 200 W, generating a core temperature of 105C at the tip. This target temperature is maintained for 10 minutes. At the completion of the ablative cycle the generator is turned off and tissue temperature is recorded at 30-second intervals. If any of the 5 thermocouple recordings document a temperature of less than 60C, the cycle is repeated. Multiple overlapping ablations are performed depending on anatomical considerations of the renal mass. At the completion of ablation the probes are retracted and the probe tract is ablated as it is removed. Spiral CT is repeated to evaluate any perinephric hematoma. RFA is generally performed on an outpatient basis.
RESULTS Table 2 lists patient demographics and pretreatment characteristics. Preoperative renal lesion characteristics were similar for RFA and cryoablation with regard to malignant or favor malignant pathology on pretreatment biopsy in 75 vs 135 patients (75% vs 70%, p ⫽ 0.33). Median renal lesion size was 2.5 vs 2.4 cm (p ⫽ 0.65). However, significantly more tumors were centrally located in the RFA group compared to those in the cryoablation series (42 or 39% vs 30 or
TABLE 1. Patient demographics and pretreatment tumor characteristics
No. pts Median age (25%–75% range) No. men (%)/women Median cm radiographic tumor size (25%–75% range) No. central (%) No. kidney status (%): Normal contralat kidney Compromised contralat kidney Solitary kidney Solitary remnant No. ablations No. pretreatment biopsy data No. benign (%) No. indeterminate (%): Favor benign Favor malignant No. malignant (%) Wilcoxon rank sum test Pearson chi-square probability
RFA
Cryoablation
88 68 (57.5–75) 74 (68)/35 (32) 2.5 (1.8–3.8) 42 (39)
176 68 (59–75) 138 (72)/54 (28) 2.4 (1.9–3.8) 30 (16)
25 35 31 18 109 99 9 35 15 20 55
(23) (32) (28) (17) (9) (35) (15) (20) (55)
92 53 46 1 192 192 47 39 29 10 106
p Value 0.9151 (Wilcoxon rank sum test) 0.4665 (Pearson chi-square probability) 0.6450 (Wilcoxon rank sum test) ⬍0.0001 (Pearson chi-square probability) ⬍0.0001 (Pearson chi-square probability)
(48) (28) (24) (0.5) (24) (20) (15) (5) (55)
0.0009 (Pearson chi-square probability)
RADIOGRAPHIC IMAGING AND HISTOPATHOLOGY OF ABLATED RENAL MASSES
1279
TABLE 2. Posttreatment success on radiographic imaging or biopsy at 6 months RFA No. 6-mo radiographic followup (%): Not enhancing Peripheral/linear enhancing Central enhancing No. 6-mo biopsy follow-up (%): Pos for atypical cells No. biopsy not completed (%) No. reason biopsy not completed (%):† Anticoagulation Lost to followup Solitary/remnant/chronic renal insufficiency Recurrence/metastatic disease Benign pretreatment biopsy Death before 6 mos No. radiographic success/total No. with 6-mo followup (%)‡ No. radiographic success with biopsy & 6-mo imaging/total No. (%) No. pathological success (%)§
73 52 10 10 37 13 72
Cryoablation
(67) (72) (14) (14) (34) (35) (66)
139 91 35 14 97 6 95
25 (35) 8 (11) 44 (61) 12 (17) 21 (29) 2 (2.8) 62/73 (85) 29/36 (81) 24 (65)
p Value*
(72) (65) (25) (10) (51) (6.2) (49)
0.2204 0.2830 0.0054 ⬍0.0001 0.0054
3 (3.2) 23 (24) 38 (41) 11 (12) 41 (43) 0 125/139 (90) 86/97 (89) 91 (94)
⬍0.0001 0.0311 0.0069 0.3447 0.0082 0.1844 0.6183 0.2248 ⬍0.0001
* Based on Pearson chi-square probability except Fisher’s exact test used for death before 6 months. † There may have been more than 1 reason. ‡ No central or nodular enhancement. § No malignant cells on 6-month biopsy.
16%, p ⬍0.0001). There were more indeterminate pathological results on pre-ablation biopsy in the RFA group (35 or 35% vs 39 or 20%, p 0.0052). Radiographic Results Posttreatment radiographic followup at 6 months was available in 73 (67%) and 139 patients (72%) who underwent RFA and cryoablation, respectively (table 1). The rate of radiographic success, defined as a lack of central or nodular enhancement on post-contrast CT or subtraction imaging MRI, was 85% (62 cases) for RFA and 90% (125) for cryoablation at 6 months of followup. Pathological Results Pathological examination after renal biopsy or subsequent radical nephrectomy was available in 37 (34%) and 97 cases (51%) for RFA and cryoablation, respectively. Most specimens were obtained from percutaneous kidney biopsies of the ablation site at 6 months in the RFA and cryoablation groups (34 or 92% and 95 or 98%, respectively). The remaining pathological specimens were obtained by radical nephrectomy spurred by obvious radiographic persistence of contrast enhancement and/or growth of the ablated lesion. All 4 of these specimens demonstrated RCC on pathological examination. Post-ablation biopsy was not done in 72 cases (69%) in the RFA group and in 95 (49%) in the cryoablation group (p ⫽ 0.0054). Table 1 lists the reasons for not performing biopsy. They varied significantly between the groups. The rate of pathological success, defined as the lack of malignant/atypical cells on post-ablation biopsy or radical nephrectomy, for RFA and cryoablation was 65% and 94% (24 and 91 patients, respectively, p ⫽ ⬍0.0001). Eight of the 10 cases of positive biopsies in the RFA cohort were retreated with percutaneous RFA. Six patients demonstrated no evidence of enhancement on subsequent followup imaging. The remaining 2 re-ablated lesions showed a persistent thin rim of enhancement and they are being followed. The other 2 positive biopsy lesions are being followed with serial imaging, which has revealed stable central enhancing lesions. Two of the 4 patients with positive biopsies in the cryoablation cohort underwent percutaneous RFA, 1
underwent repeat laparoscopic cryoablation and 1 underwent laparoscopic radical nephrectomy, which demonstrated RCC clear cell variant grade 3. In the RFA group 12 of the 13 patients with positive pathological results after ablation had a malignant (8 or 62%) or favor malignant (4 or 31%) diagnosis on pretreatment pathological assessment. However, 1 patient showed benign pathological findings on pretreatment biopsy and malignant pathological findings on followup biopsy. In the cryoablation group only 3 of 6 patients (50%) with pathologically proven RCC after ablation had a malignant (1 or 33%) or favor malignant (2 or 67%) diagnosis on pretreatment pathological evaluation. Correlation of Pathological and Radiographic Results Table 3 shows correlated radiographic and pathological data. For correlative purposes all patients except 1 underwent radiographic imaging with contrast material just before biopsy, including 36 for RFA and 97 for cryoablation. When stratified according to imaging, 6 patients (24%) without any evidence of post-ablation enhancement on 6-month biopsy were found to have viable renal cancer cells in the
TABLE 3. Contrast radiographic and pathological data No. RFA (%)
No. Cryoablation (%)
Overall No enhancement:
36 25 (69.4)
97 60 (61.9)
Neg biopsy Pos biopsy Pos peripheral enhancement: Neg biopsy
19 (76) 6 (24) 4 (11.1)
60 (100) 0 26 (26.8)
(50)
24 (92.3)
Pos biopsy Pos central enhancement: Neg biopsy
2 (50) 7 (19.4)
2 (7.7) 11 (11.3)
Pos biopsy
2
2
(29)
7
(64)
5
(71)
4
(36)
p Value 0.4181 (Pearson chi-square probability) 0.0004 (Fisher’s exact test) 0.0544 (Pearson chi-square probability) 0.0205 (Pearson chi-square probability) 0.2248 (Pearson chi-square probability) 0.1469 (Pearson chi-square probability)
1280
RADIOGRAPHIC IMAGING AND HISTOPATHOLOGY OF ABLATED RENAL MASSES
RFA group. This finding was not observed in the cryoablation group since all 60 patients (100%) without evidence of post-ablation enhancement had negative biopsies. In those with peripheral linear enhancement 2 biopsies (50%) were positive in the RFA cohort compared to only 2 (7.7%) in the cryoablation group (p ⫽ 0.0205). Centrally enhancing tumors yielded positive biopsies in 71% (5 cases) with RFA treatment but only in 36% (4) in the cryoablation group (p ⫽ 0.1469). Six of all positive post-ablation biopsies (46%) in the RFA group showed no enhancement on immediate pre-biopsy imaging, whereas all 6 positive biopsies (100%) in the post-cryoablation cohort came from tumors demonstrating some degree of enhancement. The figure shows an example of a lesion showing no enhancement at 6 months with a positive biopsy for renal cell cancer. There were 2 central enhancing lesions at 6-month followup with a negative post-ablation biopsy in the RFA group (table 3). Enhancement disappeared in 1 patient at 1-year followup and the other is being watched according to patient preference. There were 7 central enhancing lesions at 6-month followup with a negative post-ablation biopsy in the cryoablation group. Three of these patients subsequently demonstrated RCC on repeat biopsy (2) or radical nephrectomy (1). Three other patients are being watched according to patient preference and 1 was lost to followup. There were 2 peripheral enhancing lesions at 6-month followup and a negative post-ablation biopsy in the RFA group. One lesion no longer enhances and the other continues to show stable peripheral enhancement. In the cryoablation group there were 24 lesions that had thin peripheral enhancement at 6-month followup. Of these lesions 15 no longer enhanced on followup imaging, 5 were lost to followup and 2 are being watched with stable peripheral enhancement. Two patients showed radiographic recurrence, defined as a nodular central enhancing lesion. The sensitivity of central nodular enhancement on 6-month followup CT or MRI to predict a positive biopsy in our post-cryoablation cohort was 77.8% with 95.1% specificity, 63.4% PPV and 97.7% NPV. The sensitivity of central nodular enhancement on 6-month followup CT or MRI to detect a positive biopsy in our post-RFA cohort was only 38.4% with 91.3% specificity, 71.4% PPV and 72.4% NPV. DISCUSSION To our knowledge our results are unique because they represent the largest experience available for the correlation between radiographic findings and pathology results after
Example of enhancing renal tumor (A) ablated with RFA and MRI at 6 months shows no enhancement (B). Ablation bed was then biopsied and showed viable renal cell carcinoma.
ablative procedures. Traditionally ablative techniques have been considered successful if followup imaging demonstrated no enhancement or enlargement of the tumor ablation site. MRI has been cited as 99% sensitive for detecting malignant renal lesions with a PPV of 89% and NPV of 91% in the preoperative setting.9 There have been minimal data on the efficacy of MRI or CT to detect tumor destruction in post-ablative cases. Many groups have unquestioningly extrapolated the excellent diagnostic accuracy of CT and MRI to be used as the measurement of success in a post-ablative case. Rendon et al suggested caution when interpreting postablation radiographic imaging when they found that 2 of 6 patients had renal cell cancer at extirpative surgery 1 week later.11 However, this study was criticized for using dry RFA, which creates skip lesions,13 and they did not use NADH staining, which is better than hematoxylin and eosin staining for distinguishing actual viable RCC cells from ablated RCC cells in the early post-ablative setting.14,15 In our study we used the newer StarBurst FLEX RFA probe and performed followup biopsies 6 months after ablation, allowing maturation of the ablation effect and, thus, avoiding the discrepancy between hematoxylin and eosin, and NADH viability stains. Our study addressed each of these criticisms and it appears that caution is still warranted, particularly in the radiographic followup of patients treated with RFA. In the setting of the post-cryoablation patient radiographic and pathological findings seem to correlate adequately. The high NPV demonstrates that we already suspected recurrence by radiographic information in most of these patients and biopsy added little to our treatment algorithm. In contradistinction, the correlation of radiographic and pathological findings in the setting of the post-RFA patient is poor. Our radiographic success rate of 86.1% in the percutaneous RFA cohort is similar to that in other published series of comparable size and central location.16 –20 What is particularly concerning is that almost a quarter of our radiographic successes (6) actually showed a positive biopsy at 6 months. Although our methods should be reproduced at other centers to determine whether this effect and the rate at which it occurs is consistent, there may be a substantial population of patients who are cured by radiographic criteria who have viable RCC cells persisting in the kidney. However, to our knowledge the clinical significance of these viable cells is unknown at this time and longer followup is needed to determine whether these patients are at higher risk for local or systemic progression. Although biopsy added a significant amount of information it, too, has shortcomings. We observed a total of 4 cases in the 2 groups of patients in which initial pretreatment biopsy demonstrated benign pathological results and 6-month posttreatment biopsy revealed malignancy. Renal biopsy alone continues to lack a sufficient NPV to make sound oncological treatment decisions. Our study is limited by the fact that it is retrospective and only about half of the patients underwent posttreatment biopsy. It is also limited by the lack of viability staining of the pathological specimens. As mentioned, analysis of postRFA pathological assessment can be misleading using hematoxylin and eosin staining only. It appears that this difficultly may largely decrease as time from ablation increases. Martin et al found that most patients with positive hematoxylin and eosin staining at 6 months also had positive viability
RADIOGRAPHIC IMAGING AND HISTOPATHOLOGY OF ABLATED RENAL MASSES stains in a hepatocellular carcinoma model.21 Further research addressing these issues is imperative.
11.
CONCLUSIONS
12.
We observed a poor correlation between post-RFA imaging and post-RFA biopsy results at 6 months. To our knowledge the significance of these viable cells on biopsy remains to be determined. We propose that biopsy of the ablated site should be added to RFA followup protocols regardless of pretreatment biopsy results, or the presence or absence of enhancement. In contrast, lesions treated with cryotherapy seem to correlate adequately with radiographic findings and biopsy may not add substantial information.
Abbreviations and Acronyms CT ⫽ computerized tomography MRI ⫽ magnetic resonance imaging NADH ⫽ nicotinamide adenine dinucleotide (reduced form) NPV ⫽ negative predictive value PPV ⫽ positive predictive value RCC ⫽ renal cell carcinoma RFA ⫽ radio frequency ablation
13.
14.
15.
16.
17.
18.
REFERENCES 1.
2.
3.
4. 5.
6.
7.
8.
9.
10.
Frank I, Blute ML, Leibovich BC, Cheville JC, Lohse CM and Zincke H: Independent validation of the 2002 American Joint Committee on cancer primary tumor classification for renal cell carcinoma using a large, single institution cohort. J Urol 2005; 173: 1889. Fergany AF, Hafez KS and Novick AC: Long-term results of nephron sparing surgery for localized renal cell carcinoma: 10-year followup. J Urol 2000; 163: 442. Gill IS, Desai MM, Kaouk JH, Meraney AM, Murphy DP, Sung GT et al: Laparoscopic partial nephrectomy for renal tumor: duplicating open surgical techniques. J Urol 2002; 167: 469. Lane BR and Gill IS: 5-Year outcomes of laparoscopic partial nephrectomy. J Urol 2007; 177: 70. Permpongkosol S, Bagga HS, Romero FR, Sroka M, Jarrett TW and Kavoussi LR: Laparoscopic versus open partial nephrectomy for the treatment of pathological T1N0M0 renal cell carcinoma: a 5-year survival rate. J Urol 2006; 176: 1984. Ramani AP, Desai MM, Steinberg AP, Ng CS, Abreu SC, Kaouk JH et al: Complications of laparoscopic partial nephrectomy in 200 cases. J Urol 2005; 173: 42. Weld KJ, Venkatesh R, Huang J and Landman J: Evolution of surgical technique and patient outcomes for laparoscopic partial nephrectomy. Urology 2006; 67: 502. Hecht EM, Israel GM, Krinsky GA, Hahn WY, Kim DC, Belitskaya-Levy I et al: Renal masses: quantitative analysis of enhancement with signal intensity measurements versus qualitative analysis of enhancement with image subtraction for diagnosing malignancy at MR imaging. Radiology 2004; 232: 373. Hallscheidt PJ, Bock M, Riedasch G, Zuna I, Schoenberg SO, Autschbach F et al: Diagnostic accuracy of staging renal cell carcinomas using multidetector-row computed tomography and magnetic resonance imaging: a prospective study with histopathologic correlation. J Comput Assist Tomogr 2004; 28: 333. Park S, Strup SE, Saboorian H and Cadeddu JA: No evidence of disease after radiofrequency ablation in delayed nephrectomy specimens. Urology 2006; 68: 964.
19.
20.
21.
1281
Rendon RA, Kachura JR, Sweet JM, Gertner MR, Sherar MD, Robinette M et al: The uncertainty of radio frequency treatment of renal cell carcinoma: findings at immediate and delayed nephrectomy. J Urol 2002; 167: 1587. Hegarty NJ, Gill IS, Desai MM, Remer EM, O’Malley CM and Kaouk JH: Probe-ablative nephron-sparing surgery: cryoablation versus radiofrequency ablation. Urology 2006; 68: 7. Rehman J, Landman J and Sundaram CP: Re: the uncertainty of radio frequency treatment of renal cell carcinoma: findings at immediate and delayed nephrectomy. J Urol 2002; 168: 2128. Ogan K and Cadeddu JA: Re: the uncertainty of radio frequency treatment of renal cell carcinoma: findings at immediate and delayed nephrectomy. J Urol 2002; 168: 2129. Crowley JD, Shelton J, Iverson AJ, Burton MP, Dalrymple NC and Bishoff JT: Laparoscopic and computed tomographyguided percutaneous radiofrequency ablation of renal tissue: acute and chronic effects in an animal model. Urology 2001; 57: 976. Gervais DA, McGovern FJ, Arellano RS, McDougal WS and Mueller PR: Renal cell carcinoma: clinical experience and technical success with radio-frequency ablation of 42 tumors. Radiology 2003; 226: 417. Farrell MA, Charboneau WJ, DiMarco DS, Chow GK, Zincke H, Callstrom MR et al: Imaging-guided radiofrequency ablation of solid renal tumors. AJR Am J Roentgenol 2003; 180: 1509. Mayo-Smith WW, Dupuy DE, Parikh PM, Pezzullo JA and Cronan JJ: Imaging-guided percutaneous radiofrequency ablation of solid renal masses: techniques and outcomes of 38 treatment sessions in 32 consecutive patients. AJR Am J Roentgenol 2003; 180: 1503. Su LM, Jarrett TW, Chan DY, Kavoussi LR and Solomon SB: Percutaneous computed tomography-guided radiofrequency ablation of renal masses in high surgical risk patients: preliminary results. Urology 2003; 61: 26. Zagoria RJ, Hawkins AD, Clark PE, Hall MC, Matlaga BR, Dyer RB et al: Percutaneous CT-guided radiofrequency ablation of renal neoplasms: factors influencing success. AJR Am J Roentgenol 2004; 183: 201. Martin AP, Goldstein RM, Dempster J, Netto GJ, Katabi N, Derrick HC et al: Radiofrequency thermal ablation of hepatocellular carcinoma before liver transplantation–a clinical and histological examination. Clin Transplant 2006; 20: 695.
EDITORIAL COMMENT These authors evaluated the reliability of cross-sectional imaging to follow post-ablation renal tumors. To this end I believe that MRI may be less reliable than CT for assessing the adequacy of complete ablation of central tumors after RFA. However, I believe that this is a limitation of the imaging technology and not of the ablation technology. Most clinicians would agree that MRI does not distinguish tumor margins or enhancement as well as CT for most renal tumors. Unfortunately articles such as this may incorrectly provide support for those who advocate the superiority of cryoablation over RFA. As such, I believe that it important to address several methodological issues before any comparison is made. While the authors compared percutaneous RFA to laparoscopic cryoablation, there appears to have been a significant selection bias against RFA. There is no doubt that incomplete ablations are to be expected at a higher frequency in a percutaneous treatment group that includes twice as many central tumors, half as many with a normal contralateral kidney and 17 times more solitary renal remnants. It seems that RFA may have been offered to patients who were not candidates for laparoscopic cryoablation. The treating physician may have been
1282
RADIOGRAPHIC IMAGING AND HISTOPATHOLOGY OF ABLATED RENAL MASSES
more judicious with needle placement and ablation size in such a percutaneous RFA group. Any imaging technology to assess residual disease in central ablations and previously operated kidneys is likely to be more difficult to interpret. Furthermore, since the percutaneous approach is easily repeatable in such high risk groups, incomplete treatment may be more frequent than with the direct vision laparoscopic approach, which cannot be repeated, and it may result in more aggressive laparoscopic ablation. Therefore, the suggestion of superiority for cryoablation based on this study design could well be unfounded. Similarly the inference that CT or MRI can accurately interpret the cryoablation of central tumors or renal remnants may not be possible due to the paucity of such patients in this cohort. As suggested by the authors, the finding of residual viable cells identified on needle biopsy 6 months after RFA has unclear clinical significance in light of good intermediate term clinical data, varying ablation techniques and longer term pathological data collected by others.1-3 Several groups have now reported RFA results at 2, 3 and 5 years of followup, of which none approach the 35% persistent disease, recurrence or metastatic progression rate reported in this study. This discrepancy may be better explained by factors such as tumor selection (as discussed previously), ablation technique (1 RFA cycle instead of 2 routine ablations) and biopsy interpretation. The issue of biopsy interpretation is most important, concerning and difficult to understand. The authors suggest that the timing of biopsy at 6 months allows maturation of the ablation effect and avoids the discrepancy between hematoxylin and eosin, and NADH viability assessments debated in the literature. How do they know this? How do they know that they saw viable cells in the 6 of 25 cases with no enhancement but a positive biopsy? RFA treated lesions are characterized by protein denaturation and crosslinking, which stabilize the tumor architecture (heat fixation, which is similar to frozen section), resulting in a chronic granulomatous response (there is no neutrophil or macrophage infiltration, as with cryoablation) and slow reabsorption, if any. Thus, interpreting the presence of preserved cells in hematoxylin and eosin stained tissue as viable tumor may be incorrect. Instead, they may represent fixed dead cells or granulomatous inflammatory cells, as reported by others (reference 10 in article). In a similar study the University of Michigan group reported that biopsies 2 months after RFA demonstrated residual cancer on hematoxylin and eosin staining in some RFA tumors but NADH staining suggested no viable cells. We have biopsied 20 post-RFA tumors that showed no enhancement on CT, representing a population size similar to that in this study of 25 cases of negative CT or MRI, more than 1 year after RFA and in each there were no visible malignant or atypical cells. Thus, it may take a while for post-RFA heat fixated cells to degenerate, which would account for the atypical cells. Although no one would argue that residual enhancement may indicate cancer, the conclusion of a significant risk of residual renal cell cancer after RFA despite nonenhancement on cross-sectional imaging is only based on 25 cases, of which 6 (24%) had atypical cells. Interestingly in the cases of positive enhancement a similar percent had negative biopsies. Do these patients have residual cancer? In light of the pathophysiological response to thermal ablation, the biopsy experience described and the difficulty with assessing true cellular viability incomplete ablation in 24% of cases has not been confirmed with nonenhancing CT or MRI
and, therefore, convincing evidence for recommending routine needle biopsy after RFA is not provided. The authors report that biopsy after RFA is unreliable. It remains that surveillance with frequent cross-sectional imaging, preferably with CT, is the standard method for following patients after ablation therapy. Needle biopsy should be used in cases of tumor bed enhancement on cross-sectional imaging. With proper patient selection (neither central tumors nor renal remnants) cryoablation and RFA can be expected to have good intermediate term disease control. Jeffrey A. Cadeddu Urology and Radiology Clinical Center for Minimally Invasive Urologic Cancer Treatment University of Texas Southwestern Medical Center at Dallas Dallas, Texas 1.
McDougal WS, Gervais DA, McGovern FJ and Mueller PR: Long-term followup of patients with renal cell carcinoma treated with radio frequency ablation with curative intent. J Urol 2005; 174: 61. 2. Park S, Anderson JK, Matsumoto ED, Lotan Y, Josephs S and Cadeddu JA: Radiofrequency ablation of renal tumors: intermediate-term results. J Endourol 2006; 20: 569. 3. Stern J, Svatek R, Park S, Hermann M, Lotan Y, Sagalowsky AI et al: Intermediate comparison of partial nephrectomy and radiofrequency ablation for clinical T1a renal tumors. BJU Int 2007; 100: 287.
REPLY BY AUTHORS We agree that limitations of imaging do not translate into limitations of ablation, and this has never been our claim. However, when evaluating a new technology, it is prudent to understand the end points used to determine success, especially for patients with potentially lethal disease. Our study raises serious doubts about the ability to determine successful RFA with imaging alone. We believe that these end points deserve research rather than assumption and postulation. Interestingly the comment provides several potential explanations as to why such discrepancies between pathological and radiographic outcomes. We made no attempt to compare the efficacy of RFA and cryoablation but focused only on radiographically successful RFA and cryoablation, and compared how imaging after each ablative modality correlated with post-ablation biopsy. Our findings suggest that post-RFA imaging does not correlate well with post-RFA biopsy. The 3 articles cited in the comment are almost entirely based on radiographic imaging, with little post-RFA pathological data. The first report had no pathological followup data (reference 1 in comment), and the other 2 reports include pathological followup data on 2 cases (references 2 and 3 in comment). We commend Cadeddu for implementing post-ablation biopsy which is exactly the data we need to be able to make informed decisions about ablative techniques. The ability to interpret H&E staining after RFA is a complicated subject. A 6-month biopsy appears to be adequate because there are data in the hepatocellular RFA literature showing that the majority of patients with positive H&E staining at 6 months will also have positive viability stains (reference 21 in article) and these positive biopsies clearly looked different from negative sites. If this was just a fixation effect,
RADIOGRAPHIC IMAGING AND HISTOPATHOLOGY OF ABLATED RENAL MASSES we would expect to see nearly all positive biopsies. Our biopsy findings were not “atypical cells with inflammatory response,” but rather textbook images of RCC. So far, only the RFA literature shows at least debatable inconsistency between H&E and NADH staining, ghost cells, and discrepancy between radiographic and pathological definitions of success. For researchers and clinicians these findings should at least be flagged and further investigated, rather than finding good scenario conclusions up front and ignoring a potential problem in the oncologic adequacy of a given technique.
1283
We agree that cryoablation and RFA for selected small and superficial renal lesions are expected to give good results. However partial nephrectomy remains the gold standard for such tumors, and we disagree that “proper patient selection” excludes central tumors and renal remnants. In fact, this subgroup of patients may benefit the most from thermal ablation when no other surgical management is feasible. Obtaining pathology data after RFA seems even more relevant given the confusion regarding interpretation of imaging as well as the biopsies. We owe our patients the diligence to sort out this issue.