- Email: [email protected]
The incidence, predictors, and survival of disappearing small renal masses on active surveillance
ARTICLE IN PRESS
Urologic Oncology: Seminars and Original Investigations 000 (2019) 1−6
Clinical-Kidney cancer
The incidence, predictors, and survival of disappearing small renal masses on active surveillance Arnav Srivastava, M.D./MPHa,*, Hiten D. Patel, MPH/M.D.a, Mohit Gupta, M.D.a, Gregory A. Joice, M.D.a, Zeyad Schwen, M.D.a, Ridwan Alam, M.D./MPHa, Michael A. Gorin, M.D.a, Michael H. Johnson, M.D.a, Bruce J. Trock, Ph.D.a, Peter Chang, M.D.b, Andrew A. Wagner, M.D.b, James M. McKiernan, M.D.c, Mohamad E. Allaf, M.D.a, Phillip M. Pierorazio, M.D.a a
James Buchanan Brady Urological Institute, Johns Hopkins Hospital, Baltimore, MD b Division of Urology, Beth Israel Deaconess Medical Center, Boston, MA c Department of Urology Columbia University Medical Center, New York, NY
Received 11 April 2019; received in revised form 30 August 2019; accepted 7 October 2019
Abstract Objective: To evaluate the incidence, predictors, and survival for those small renal masses (SRM, solid mass ≤4 cm suspicious for a clinical T1a renal cell carcinoma) that disappear on imaging while undergoing active surveillance (AS). Subjects/patients and methods: The Delayed Intervention and Surveillance for SRM registry prospectively enrolled 739 patients with SRMs. Patients having at least 1 image showing no lesion were considered to have a “disappearing” SRM. Logistic regression assessed predictors of having a disappearing SRM and Kaplan-Meier estimates illustrated relative survival. Results: Of 374 patients enrolled in AS, 22 (5.9%) experienced a disappearing SRM. Mean time to tumor disappearance was 2.0 years (SD = 1.9) and 50.0% reappeared on subsequent CT imaging. SRM disappearance, most commonly encountered on ultrasound imaging surveillance, was independently associated with tumors <1 cm on multivariable analysis (OR = 10.6 (95% CI: 1.1−100.3), P = 0.04). Furthermore, patients with disappearing SRMs were healthier than other patients on AS with no compromise in overall survival during follow-up (5-year survival = 100% vs. 73.2%, P = 0.06). Conclusions: Approximately 5% of SRM on AS will disappear during follow-up on surveillance imaging. Most of these represent artifacts of heterogeneous imaging modalities, including ultrasound, and the SRM will reappear on subsequent imaging. Given the indolent nature of these lesions, disappearance events do not require reflex repeat imaging and patients should continue AS with their original surveillance schedule intact. A smaller percentage of patients undergoing AS for a SRM may have a mass the permanently disappears. Ó 2019 Elsevier Inc. All rights reserved.
Keywords: Disappearing tumors; Kidney cancer; Active surveillance
1. Introduction While a recognized phenomenon, spontaneous clinical regression of kidney cancer is infrequently documented and its clinical implications remain poorly understood. Most reported cases of renal cell carcinoma (RCC) regression *Corresponding author. Tel.: +1-908-217-0765; fax: +1-410-502-7711. E-mail addresses: [email protected], [email protected] (A. Srivastava). https://doi.org/10.1016/j.urolonc.2019.10.005 1078-1439/Ó 2019 Elsevier Inc. All rights reserved.
describe spontaneous regression of metastases. Conversely, regression of primary renal tumors remains rarely reported within the existing literature [1]. With the growth of active surveillance (AS) for small renal masses (SRMs, solid mass ≤4 cm suspicious for cT1a RCC), regular imaging surveillance provides a safe and unique opportunity to study renal tumors that disappear on surveillance imaging. The Delayed Intervention and Surveillance for Small Renal Masses (DISSRM) Registry, a large multi-institutional prospective database, has demonstrated the safety of
ARTICLE IN PRESS 2
A. Srivastava et al. / Urologic Oncology: Seminars and Original Investigations 00 (2019) 1−6
Fig. 1. Timeline of disappearance and reappearance of renal masses on AS. AS = active surveillance.
AS in the management of SRMs [2]. Within the DISSRM Registry, we have encountered “disappearing” SRMs— those tumors with at least 1 surveillance image showing no renal lesion. These disappearance encounters include masses that disappear for varying lengths of time—including a smaller subset which may permanently disappear—and across different imaging modalities (Fig. 1). To date, however, no guidelines or data exist regarding the prognostic behavior of these tumors and appropriate management strategy. We hypothesized that imaging modality heterogeneity, particularly the use of renal ultrasound, explains many SRM “disappearance” events. Consequently, those tumors which transiently disappear would be small tumors. However, those tumors which disappear for longer periods, or even permanently disappear, may represent a low-risk subset of SRMs undergoing AS. Therefore, we aimed to evaluate the characteristics and predictors of tumors that disappear while undergoing AS for SRMs in a prospective cohort.
history, familial RCC history, or suspicion of another cancer metastasizing to the kidney were excluded. Included patients were counseled regarding the risks and benefits of primary intervention, AS, focal ablation, and percutaneous biopsy. All patients received baseline physical exam, complete blood count, comprehensive metabolic panel, and chest imaging [2]. Of the 739 prospectively enrolled patients in the DISSRM Registry, 374 elected AS. Regarding imaging surveillance, the DISSRM protocol recommends axial imaging within 6 months of AS enrollment. After this, serial imaging is performed, alternating between ultrasound and axial imaging (CT [3 or 5 mm cuts] or MRI) [2]. All patients included in analysis had axial imaging confirming the presence of a solid cortical mass at least once. We considered AS patients with at least 1 surveillance image showing no visible lesion to have a disappearing SRM. 2.2. Data collection and analysis
2. Methods 2.1. Study population and protocol The DISSRM registry is a prospective multi-institutional cohort, enrolling patients at Johns Hopkins Hospital (Baltimore, MD), Columbia University Medical Center (New York City, New York), and Beth Israel Deaconess Medical Center (Boston, MA). Patients at least 18 years old with a clinically localized, solid cortical renal lesion ≤4 cm (suspicious for cT1a RCC) were included. Those with prior RCC
For patients meeting inclusion criteria, baseline data was collected and patient outcomes / imaging results were prospectively recorded. We aimed to study incidence and prevalence of SRM disappearance, predictors of disappearance, and implications for overall survival. The proportion of patients with disappearing SRMs was tabulated against all patients undergoing AS. Additionally, point prevalence and inverse Kaplan-Meier estimates—used to approximate cumulative incidence—were calculated over 5 years. Univariable and multivariable logistic regression assessed
ARTICLE IN PRESS A. Srivastava et al. / Urologic Oncology: Seminars and Original Investigations 00 (2019) 1−6
predictors of having a disappearing SRM. Statistical associations were assessed against the whole DISSRM registry as well as those just undergoing AS. Lastly, Kaplan-Meier estimates illustrated the relative overall survival of patients with disappearing SRMs. All analyses were conducted using STATA software (v. 14.0, StataCorp, College Station, TX). 3. Results Of 374 patients electing AS, 22 (5.9%) had a disappearing SRM. Median follow-up was 3.0 years (interquartile range: [1.4−5.1 years]) and mean follow-up among disappearing SRMs and other patients on AS was similar (3.3 years vs. 2.5 years, P = 0.08). Mean tumor diameter (1.83 [range 0.6−4.3] cm vs. 1.99 [0.4−7.7] cm, P = 0.44) and crossover to primary intervention (9.1% vs. 13.6%, P = 0.54) also were similar between disappearing SRMs and other SRMs on surveillance. Notably, patients with a disappearing SRMs typically had a lower (Charleston Comorbidity Index = 0: 68.2% vs. 41.8%, P = 0.04) (Table 1). Two patients had received biopsy of their tumor prior to disappearance which demonstrated a benign mass in 1 case and a nondiagnostic result suggestive of fibrosis in the other. Two patients crossed over to intervention with 1 showing ISUP grade 2 clear cell RCC after partial nephrectomy, and 1 had no pathology due to receiving stereotactic
3
radiosurgery. Among nondisappearing SRMs, RCC (45.2%), and oncocytic (42.9%) histologies constituted most biopsied masses. Of the 22 patients with disappearing SRMs, 15 had their disappearance event (i.e. imaging showing no renal lesion) with ultrasound (68.2%), 5 (22.7%) with CT, and 2 (9.1%) with MRI. On initial imaging, upon entry into the registry, those masses disappearing on CT were 3.6 cm (ID# 1 and 21), 1.6 cm (ID# 8), 0.6 cm (ID# 14), and 1.8 m (ID# 16) (Fig. 1). On MRI disappearing masses were 1.1 cm (ID# 6) and 2.1 cm (ID#10) at initial measurement (Fig. 1). Mean time to disappearing SRM event was 2.0 years (SD = 1.9) and 50.0% of tumors reappeared on subsequent CT imaging (11 of 22). Fig. 1 further illustrates the time to disappearance and reappearance for each identified disappearing SRM. Mean growth rate for disappearing SRMs was 0.26 cm/year (SD = 1.47 cm/year) prior to disappearance. No AS patients experienced progression to metastatic kidney cancer. Disappearing SRM proportions were further examined with cumulative incidence (13.3% at 5 years, [Fig. 2]) and point prevalence (2.9% at 5 years, [Supplemental Fig. 1]). Tumor diameter <1 cm significantly predicted a disappearing SRM on univariable and multivariable analysis (Multivariable odds ratio = 10.56 (95% CI: 1.11−100.32), P = 0.04) (Table 2). No other variables significantly
Table 1 Cohort demographics by mass disappearance Active surveillance
N Follow-up (y) Age (y) BMI Sex (male) Race
Crossover to primary intervention Charleston comorbidity index
Baseline creatinine at diagnosis Mass laterality
Mass diameter at diagnosis (cm) Mass diameter ≤1.0 cm Arterial Hounsfield Units at diagnosis (HU) Incidentaloma Classic symptoms Renal biopsy of mass
White Black Asian Native American Other 0 1−3 ≥4 Left Right Bilateral
Disappearing small renal masses
Mean/count
SD/percentage
Mean/count
SD/percentage
P value
352 2.5 69.9 28.7 200 272 57 8 1 4 48 147 179 26 1.21 161 170 18 1.99 29 77.82 314 40 53
22 2.0 10.9 7.7 56.8 79.5% 16.7% 2.3% 0.3% 1.2% 13.6% 41.8% 50.9% 7.4% 85.0% 46.1% 48.7% 5.2% 0.9 8.2% 48.8 92.9% 11.4% 15.3%
3.3 68.5 27.4 12 18 3 0 0 0 2 15 7 0 1.24 8 14 0 1.83 4 79.10 19 4 2
2.4 11.6 7.2 54.6 85.7% 14.3% 0.0% 0.0% 0.0% 9.1% 68.2% 31.8% 0.0% 50.0% 36.4% 63.6% 0.0% 0.9 18.2% 49.2 90.5% 18.2% 9.1%
0.08 0.55 0.44 0.83 0.92 0.55 0.04 0.92 0.29 0.44 0.11 0.96 0.68 0.34 0.43
BMI = body mass index; SRM = small renal masses; SD = standard deviation. P < 0.05 *Classic symptoms = having at least 1 of the following symptoms: flank pain, palpable abdominal mass, hematuria.
ARTICLE IN PRESS A. Srivastava et al. / Urologic Oncology: Seminars and Original Investigations 00 (2019) 1−6
4
Fig. 2. Cumulative incidence of mass disappearance.
predicted the occurrence of a disappearing SRM. Survival analysis displayed in Supplemental Fig. 2 demonstrates noninferior overall survival among patients with disappearing SRMs compared to others undergoing AS (5-year survival: 100% vs. 73.2%, P = 0.06). 4. Discussion Our study quantifies the rates of disappearing SRMs in a prospective cohort and offers several interesting findings
regarding the management of SRMs on AS. First, 5.9% (22/374) of SRM patients undergoing AS have at least 1 surveillance image showing no lesion. Point prevalence increased from 1.3%−2.9% over 5 years. Furthermore, as follow-up accrues, 5-year cumulative incidence is 13.3%. On average disappearing SRMs had imaging showing no lesion after 2.0 years of follow-up and most commonly after ultrasound surveillance. Fifty percent of the disappearing lesions reappeared on subsequent CT imaging. When assessing statistical correlates of a disappearing SRM,
Table 2 Predictors of disappearing SRM among DISSRM registry Predictors of disappearance vs. entire registry Univariable Variable
OR
Age BMI Male Mass diameter <1 cm Mass diameter <2 cm Incidentaloma Arterial Hounsfield Units Laterality = Left
1.03 0.96 0.84 4.57 1.67 0.75 1.00 1.90
95% CI 0.99 0.91 0.37 1.47 0.74 0.17 0.98 0.80
Multivariable P value
1.07 1.02 1.91 14.19 3.79 3.28 1.02 4.50
0.12 0.19 0.69 0.01 0.22 0.70 0.68 0.14
OR 0.99 0.95 0.18 10.56 − 0.33 1.00 1.17
95% CI 0.90 0.79 0.02 1.11 − 0.02 0.97 0.15
P value 1.09 1.15 1.90 100.32 − 6.49 1.02 9.05
0.91 0.60 0.16 0.04 − 0.47 0.70 0.88
Predictors of disappearance vs. active surveillance masses only Variable
OR
95% CI
Age BMI Male Mass diameter <1 cm Mass diameter <2 cm Incidentaloma Arterial Hounsfield Units Laterality = Left
0.99 0.98 0.91 2.48 1.12 0.73 1.00 1.66
0.95 0.93 0.38 0.79 0.47 0.16 0.98 0.68
1.03 1.03 2.17 7.80 2.70 3.30 1.02 4.06
P value
OR
95% CI
0.55 0.44 0.84 0.12 0.79 0.68 0.95 0.27
0.95 0.91 0.22 2.64 − 0.45 1.00 1.30
0.84 0.74 0.02 0.27 − 0.02 0.98 0.12
95% CI = 95% confidence interval; BMI = body mass index; OR = odds ratio; SRM = small renal mass.
P value 1.06 1.12 2.62 25.88 − 9.62 1.02 14.26
0.35 0.39 0.23 0.41 − 0.61 0.92 0.83
ARTICLE IN PRESS A. Srivastava et al. / Urologic Oncology: Seminars and Original Investigations 00 (2019) 1−6
tumor size <1 cm was the strongest independent predictor. Given, the importance of tumor size and imaging modality in our analysis, we believe that many of these “disappearance” events stem from limitations of ultrasound and the use of multiple imaging modalities. Lastly, patients with disappearing SRMs did not have compromised outcomes and experienced excellent 5-year overall survival. While no prior series quantify rates of spontaneous tumor regression of RCC, several case reports exist in the current literature [1,3−7]. Based on these previous case reports, Wein et al. estimate this incidence of 0.3%−1% [8]. However our results describe a higher incidence, with an absolute proportion of 5.9% and 5-year cumulative incidence of 13.3%. Scheduled imaging surveillance in our AS protocol increases sensitivity when detecting disappearing SRMs. Additionally, regression of RCC metastases mostly constitute old estimates. A review by Jaisezewska et al. finds that 81.3% (48/59) of prior spontaneous RCC regression reports describe regression of metastases as opposed to 8.5% (5/59) describing primary tumor regression. Differently, our study emphasizes variation in imaging detection or possible clinical regression of primary localized tumors among low risk SRMs amenable to AS [3]. Over half of disappearing SRMs reappear on later imaging. In these cases, “disappearance” does not necessarily represent spontaneous tumor regression, but likely spatial and technical limitations of surveillance imaging—particularly renal ultrasound. However, as most tumors on AS are small cT1a lesions, risk of pT3a upstaging or development of metastatic disease is minimal [9,10]. Thus, while ultrasound may have inferior evaluation of the regional lymph nodes and extra renal fat invasion, it safely reduces cost and radiation for those patients on AS. Furthermore, previous studies demonstrate similar renal mass size estimates between ultrasound, CT, and MRI [11]. However, for very small SRMs, which stress the boundaries of imaging resolution, the user-dependent nature of ultrasound may result in more false negative imaging than axial imaging. Also patients with renal medullary necrosis may have obscuring of their renal mass, even on axial imaging delayed contrast phases. Consequently, most disappearing SRMs were encountered on ultrasound and most disappearing masses were <1 cm. Our study may also support that this low risk subset of patients, may actually require less rigorous imaging, than those with larger masses. Interestingly, while not statistically significant, those patients with disappearing masses trended towards having longer follow-up (3.3 vs. 2.5 years, P = 0.08). This may represent patient concerns regarding their mass and clinician efforts to make sure the mass the is recharacterized on subsequent imaging. Therefore, we reassure patients that their SRM may not be “gone” and inform them reappearance is likely on subsequent imaging. Understanding that patients with disappearing SRMs do not experience poor outcomes, we recommend that urologists not follow-up tumor disappearance with immediate reflex imaging. Instead, patients should continue
5
on their original surveillance schedule, sparing them additional radiation and healthcare costs. Finally, our study provides further evidence that AS is safe for SRMs, especially those <1 cm in diameter. 11.4% (4/35) of tumors <1 cm in the DISSRM registry had negative surveillance imaging at least once. Most importantly, disappearing SRMs demonstrate favorable tumor biology and behavior. Not requiring biopsy at enrollment, the DISSRM registry includes patients with malignant and benign tumors. Given the diagnostic limitations of imaging and renal mass biopsy, particularly for SRMs ≤1 cm, this strategy continues to prove safe. For disappearing SRMs, linear growth rates are slow, no patient has developed metastatic disease, and 5-year survival is 100%. SRMs that truly disappear may represent spontaneously resolving benign processes (i.e. infectious or inflammatory) or indolent RCC regressing due to immune surveillance. The improved overall survival reflects good patient health and well-functioning immune systems. A large population-based study examining perioperative infection after nephrectomy hypothesizes that immunologic defenses and prevention of T cell anergy may improve RCC outcomes [12]. Similarly, low risk renal masses may permanently regress due to up-regulation of inflammatory cytokines and cytotoxic T cells. However, we note that heterogenous surveillance imaging, difficult to characterize masses, and limitations in spatial resolution may primarily drive many of these “disappearance" events. Several potential limitations should be considered for the present report. First, the overall proportion of disappearing SRMs remains low, under 6%—but is significantly higher in patients with SRMs <1 cm. Furthermore, as previously described, many of these “disappearing” SRMs occur when using ultrasound and represent limitations in current imaging. With future imaging and followup, these lesions may reappear and depress the incidence we have calculated. Also, sample size may limit conclusions regarding incidence and relative survival of patients with disappearing SRMs. While additional pathological data would be informative, this report demonstrates that very small SRMs, of 1 and 2 cm, had the highest likelihood of “disappearance.” Lastly, we utilized overall survival as a measure of outcomes amongst patients with disappearing masses. Given how few patients truly die from their SRM, it is unlikely that our analysis would uncover a large difference in survival. However, it does allow us to demonstrate that those patients with disappearing SRMs have not been managed unsafely and do not require increased imaging surveillance. This supports recent guidelines for AS as initial management of patients with SRMs ≤2 cm [13]. Our report provides additional evidence regarding the indolent nature of these tumors, questions the ability of current imaging techniques to accurately discriminate very small SRMs, and informs patients and providers that a small, but significant proportion of SRMs may clinically regress.
ARTICLE IN PRESS A. Srivastava et al. / Urologic Oncology: Seminars and Original Investigations 00 (2019) 1−6
6
5. Conclusion AS for SRMs ≤2 cm does not compromise oncologic outcomes for patients with suspected RCC. Approximately 5% of SRMs on AS may have their lesion disappear on surveillance imaging, particularly after ultrasound and in those tumors <1 cm in diameter. Many tumors reappear on subsequent imaging indicating limitations in spatial imaging and potential user dependence. However, patients with disappearing SRMs do not experience compromised outcomes, and a proportion may spontaneously regress. Consequently, disappearing SRMs do not require reflex axial imaging and patients can continue on their original imaging schedule. Supplementary materials Supplementary material associated with this article can be found in the online version at https://doi.org/10.1016/j. urolonc.2019.10.005. References [1] Jawanda GG, Drachenberg D. Spontaneous regression of biopsy proven primary renal cell carcinoma: a case study. Can Urol Assoc 2012;6(5):E203–5. https://doi.org/10.5489/cuaj.11035. [2] Pierorazio PM, Johnson MH, Ball MW, Gorin MA, Trock BJ, Chang P, et al. Five-year analysis of a multi-institutional prospective clinical trial of delayed intervention and surveillance for small renal masses: the DISSRM registry. Eur Urol 2015;68(3):408–15. https://doi.org/ 10.1016/j.eururo.2015.02.001. [3] Janiszewska AD, Poletajew S, Wasiutynski A. Reviews spontaneous regression of renal cell carcinoma. Wspo»czesna Onkol 2013;2(2): 123–7. https://doi.org/10.5114/wo.2013.34613.
[4] Snow RM, Schellhammer PF. Spontaneous regression of metastatic renal cell carcinoma. Urology 1982;20(2):177–81. https://doi.org/ 10.1016/0090-4295(82)90356-9. [5] Oliver RTD, Nethersell ABW, Bottomley JM. Unexplained spontaneous regression and alpha interferon as treatment for metastatic renal carcinoma. Br J Urol 1989;63(2):128–31. https://doi.org/10.1111/ j.1464-410X.1989.tb05147.x. [6] de Riese W, Goldenberg K, Allhoff E, Stief C, Schlick R, Liedke S, et al. Metastatic renal cell carcinoma (RCC): spontaneous regression, long-term survival and late recurrence. Int Urol Nephrol 1991; 23(1):13–25: Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/ 1938215. [7] Srivastava A, Meyer AR, Pierorazio PM, Rowe SP, Allaf ME, Gorin MA. Spontaneous regression of a low-grade renal cell carcinoma with oncocytic features after renal mass biopsy. Clin Genitourin Cancer 2018. https://doi.org/10.1016/j.clgc.2018.07.012. [8] Wein AJ, Kavoussi LR, Novick AC, Partin AW 2012, Peters Ca. Campbell-Walsh Urology. 10th Edition. Campbell-Walsh Urology. https://doi.org/10.1016/B978-1-4160-6911-9.00061-X. [9] Srivastava A, Patel HD, Joice GA, Semerjian A, Gorin MA, Johnson MH, et al. Incidence of T3a up-staging and survival after partial nephrectomy: size-stratified rates and implications for prognosis. Urol Oncol 2018;36(1):7–13. https://doi.org/10.1016/j.urolonc.2017. 09.005. [10] Smaldone MC, Kutikov A, Egleston BL, Canter DJ, Viterbo R, Chen DYT, et al. Small renal masses progressing to metastases under active surveillance. Cancer 2012;118(4):99–1006. https://doi.org/10.1002/ cncr.26369. [11] Mucksavage P, Ramchandani P, Malkowicz SB, Guzzo TJ. Is ultrasound imaging inferior to computed tomography or magnetic resonance imaging in evaluating renal mass zize? Urology 2012;79(1): 28–31. https://doi.org/10.1016/J.UROLOGY.2011.09.036. [12] Tallman JE, Pearce SM, Kuchta K, Helfand BT, Eggener SE. Impact of perioperative infection on cancer specific survival after nephrectomy for renal cell carcinoma. J Urol 2017;198(5):1027–32. https:// doi.org/10.1016/J.JURO.2017.05.070. [13] Campbell S, Uzzo RG, Allaf ME, Bass EB, Cadeddu JA, Chang A, et al. Renal mass and localized renal cancer: AUA guideline. J Urol 2017;198(3):520–9. https://doi.org/10.1016/j.juro.2017.04.100.
Urologic Oncology: Seminars and Original Investigations 000 (2019) 1−6
Clinical-Kidney cancer
The incidence, predictors, and survival of disappearing small renal masses on active surveillance Arnav Srivastava, M.D./MPHa,*, Hiten D. Patel, MPH/M.D.a, Mohit Gupta, M.D.a, Gregory A. Joice, M.D.a, Zeyad Schwen, M.D.a, Ridwan Alam, M.D./MPHa, Michael A. Gorin, M.D.a, Michael H. Johnson, M.D.a, Bruce J. Trock, Ph.D.a, Peter Chang, M.D.b, Andrew A. Wagner, M.D.b, James M. McKiernan, M.D.c, Mohamad E. Allaf, M.D.a, Phillip M. Pierorazio, M.D.a a
James Buchanan Brady Urological Institute, Johns Hopkins Hospital, Baltimore, MD b Division of Urology, Beth Israel Deaconess Medical Center, Boston, MA c Department of Urology Columbia University Medical Center, New York, NY
Received 11 April 2019; received in revised form 30 August 2019; accepted 7 October 2019
Abstract Objective: To evaluate the incidence, predictors, and survival for those small renal masses (SRM, solid mass ≤4 cm suspicious for a clinical T1a renal cell carcinoma) that disappear on imaging while undergoing active surveillance (AS). Subjects/patients and methods: The Delayed Intervention and Surveillance for SRM registry prospectively enrolled 739 patients with SRMs. Patients having at least 1 image showing no lesion were considered to have a “disappearing” SRM. Logistic regression assessed predictors of having a disappearing SRM and Kaplan-Meier estimates illustrated relative survival. Results: Of 374 patients enrolled in AS, 22 (5.9%) experienced a disappearing SRM. Mean time to tumor disappearance was 2.0 years (SD = 1.9) and 50.0% reappeared on subsequent CT imaging. SRM disappearance, most commonly encountered on ultrasound imaging surveillance, was independently associated with tumors <1 cm on multivariable analysis (OR = 10.6 (95% CI: 1.1−100.3), P = 0.04). Furthermore, patients with disappearing SRMs were healthier than other patients on AS with no compromise in overall survival during follow-up (5-year survival = 100% vs. 73.2%, P = 0.06). Conclusions: Approximately 5% of SRM on AS will disappear during follow-up on surveillance imaging. Most of these represent artifacts of heterogeneous imaging modalities, including ultrasound, and the SRM will reappear on subsequent imaging. Given the indolent nature of these lesions, disappearance events do not require reflex repeat imaging and patients should continue AS with their original surveillance schedule intact. A smaller percentage of patients undergoing AS for a SRM may have a mass the permanently disappears. Ó 2019 Elsevier Inc. All rights reserved.
Keywords: Disappearing tumors; Kidney cancer; Active surveillance
1. Introduction While a recognized phenomenon, spontaneous clinical regression of kidney cancer is infrequently documented and its clinical implications remain poorly understood. Most reported cases of renal cell carcinoma (RCC) regression *Corresponding author. Tel.: +1-908-217-0765; fax: +1-410-502-7711. E-mail addresses: [email protected], [email protected] (A. Srivastava). https://doi.org/10.1016/j.urolonc.2019.10.005 1078-1439/Ó 2019 Elsevier Inc. All rights reserved.
describe spontaneous regression of metastases. Conversely, regression of primary renal tumors remains rarely reported within the existing literature [1]. With the growth of active surveillance (AS) for small renal masses (SRMs, solid mass ≤4 cm suspicious for cT1a RCC), regular imaging surveillance provides a safe and unique opportunity to study renal tumors that disappear on surveillance imaging. The Delayed Intervention and Surveillance for Small Renal Masses (DISSRM) Registry, a large multi-institutional prospective database, has demonstrated the safety of
ARTICLE IN PRESS 2
A. Srivastava et al. / Urologic Oncology: Seminars and Original Investigations 00 (2019) 1−6
Fig. 1. Timeline of disappearance and reappearance of renal masses on AS. AS = active surveillance.
AS in the management of SRMs [2]. Within the DISSRM Registry, we have encountered “disappearing” SRMs— those tumors with at least 1 surveillance image showing no renal lesion. These disappearance encounters include masses that disappear for varying lengths of time—including a smaller subset which may permanently disappear—and across different imaging modalities (Fig. 1). To date, however, no guidelines or data exist regarding the prognostic behavior of these tumors and appropriate management strategy. We hypothesized that imaging modality heterogeneity, particularly the use of renal ultrasound, explains many SRM “disappearance” events. Consequently, those tumors which transiently disappear would be small tumors. However, those tumors which disappear for longer periods, or even permanently disappear, may represent a low-risk subset of SRMs undergoing AS. Therefore, we aimed to evaluate the characteristics and predictors of tumors that disappear while undergoing AS for SRMs in a prospective cohort.
history, familial RCC history, or suspicion of another cancer metastasizing to the kidney were excluded. Included patients were counseled regarding the risks and benefits of primary intervention, AS, focal ablation, and percutaneous biopsy. All patients received baseline physical exam, complete blood count, comprehensive metabolic panel, and chest imaging [2]. Of the 739 prospectively enrolled patients in the DISSRM Registry, 374 elected AS. Regarding imaging surveillance, the DISSRM protocol recommends axial imaging within 6 months of AS enrollment. After this, serial imaging is performed, alternating between ultrasound and axial imaging (CT [3 or 5 mm cuts] or MRI) [2]. All patients included in analysis had axial imaging confirming the presence of a solid cortical mass at least once. We considered AS patients with at least 1 surveillance image showing no visible lesion to have a disappearing SRM. 2.2. Data collection and analysis
2. Methods 2.1. Study population and protocol The DISSRM registry is a prospective multi-institutional cohort, enrolling patients at Johns Hopkins Hospital (Baltimore, MD), Columbia University Medical Center (New York City, New York), and Beth Israel Deaconess Medical Center (Boston, MA). Patients at least 18 years old with a clinically localized, solid cortical renal lesion ≤4 cm (suspicious for cT1a RCC) were included. Those with prior RCC
For patients meeting inclusion criteria, baseline data was collected and patient outcomes / imaging results were prospectively recorded. We aimed to study incidence and prevalence of SRM disappearance, predictors of disappearance, and implications for overall survival. The proportion of patients with disappearing SRMs was tabulated against all patients undergoing AS. Additionally, point prevalence and inverse Kaplan-Meier estimates—used to approximate cumulative incidence—were calculated over 5 years. Univariable and multivariable logistic regression assessed
ARTICLE IN PRESS A. Srivastava et al. / Urologic Oncology: Seminars and Original Investigations 00 (2019) 1−6
predictors of having a disappearing SRM. Statistical associations were assessed against the whole DISSRM registry as well as those just undergoing AS. Lastly, Kaplan-Meier estimates illustrated the relative overall survival of patients with disappearing SRMs. All analyses were conducted using STATA software (v. 14.0, StataCorp, College Station, TX). 3. Results Of 374 patients electing AS, 22 (5.9%) had a disappearing SRM. Median follow-up was 3.0 years (interquartile range: [1.4−5.1 years]) and mean follow-up among disappearing SRMs and other patients on AS was similar (3.3 years vs. 2.5 years, P = 0.08). Mean tumor diameter (1.83 [range 0.6−4.3] cm vs. 1.99 [0.4−7.7] cm, P = 0.44) and crossover to primary intervention (9.1% vs. 13.6%, P = 0.54) also were similar between disappearing SRMs and other SRMs on surveillance. Notably, patients with a disappearing SRMs typically had a lower (Charleston Comorbidity Index = 0: 68.2% vs. 41.8%, P = 0.04) (Table 1). Two patients had received biopsy of their tumor prior to disappearance which demonstrated a benign mass in 1 case and a nondiagnostic result suggestive of fibrosis in the other. Two patients crossed over to intervention with 1 showing ISUP grade 2 clear cell RCC after partial nephrectomy, and 1 had no pathology due to receiving stereotactic
3
radiosurgery. Among nondisappearing SRMs, RCC (45.2%), and oncocytic (42.9%) histologies constituted most biopsied masses. Of the 22 patients with disappearing SRMs, 15 had their disappearance event (i.e. imaging showing no renal lesion) with ultrasound (68.2%), 5 (22.7%) with CT, and 2 (9.1%) with MRI. On initial imaging, upon entry into the registry, those masses disappearing on CT were 3.6 cm (ID# 1 and 21), 1.6 cm (ID# 8), 0.6 cm (ID# 14), and 1.8 m (ID# 16) (Fig. 1). On MRI disappearing masses were 1.1 cm (ID# 6) and 2.1 cm (ID#10) at initial measurement (Fig. 1). Mean time to disappearing SRM event was 2.0 years (SD = 1.9) and 50.0% of tumors reappeared on subsequent CT imaging (11 of 22). Fig. 1 further illustrates the time to disappearance and reappearance for each identified disappearing SRM. Mean growth rate for disappearing SRMs was 0.26 cm/year (SD = 1.47 cm/year) prior to disappearance. No AS patients experienced progression to metastatic kidney cancer. Disappearing SRM proportions were further examined with cumulative incidence (13.3% at 5 years, [Fig. 2]) and point prevalence (2.9% at 5 years, [Supplemental Fig. 1]). Tumor diameter <1 cm significantly predicted a disappearing SRM on univariable and multivariable analysis (Multivariable odds ratio = 10.56 (95% CI: 1.11−100.32), P = 0.04) (Table 2). No other variables significantly
Table 1 Cohort demographics by mass disappearance Active surveillance
N Follow-up (y) Age (y) BMI Sex (male) Race
Crossover to primary intervention Charleston comorbidity index
Baseline creatinine at diagnosis Mass laterality
Mass diameter at diagnosis (cm) Mass diameter ≤1.0 cm Arterial Hounsfield Units at diagnosis (HU) Incidentaloma Classic symptoms Renal biopsy of mass
White Black Asian Native American Other 0 1−3 ≥4 Left Right Bilateral
Disappearing small renal masses
Mean/count
SD/percentage
Mean/count
SD/percentage
P value
352 2.5 69.9 28.7 200 272 57 8 1 4 48 147 179 26 1.21 161 170 18 1.99 29 77.82 314 40 53
22 2.0 10.9 7.7 56.8 79.5% 16.7% 2.3% 0.3% 1.2% 13.6% 41.8% 50.9% 7.4% 85.0% 46.1% 48.7% 5.2% 0.9 8.2% 48.8 92.9% 11.4% 15.3%
3.3 68.5 27.4 12 18 3 0 0 0 2 15 7 0 1.24 8 14 0 1.83 4 79.10 19 4 2
2.4 11.6 7.2 54.6 85.7% 14.3% 0.0% 0.0% 0.0% 9.1% 68.2% 31.8% 0.0% 50.0% 36.4% 63.6% 0.0% 0.9 18.2% 49.2 90.5% 18.2% 9.1%
0.08 0.55 0.44 0.83 0.92 0.55 0.04 0.92 0.29 0.44 0.11 0.96 0.68 0.34 0.43
BMI = body mass index; SRM = small renal masses; SD = standard deviation. P < 0.05 *Classic symptoms = having at least 1 of the following symptoms: flank pain, palpable abdominal mass, hematuria.
ARTICLE IN PRESS A. Srivastava et al. / Urologic Oncology: Seminars and Original Investigations 00 (2019) 1−6
4
Fig. 2. Cumulative incidence of mass disappearance.
predicted the occurrence of a disappearing SRM. Survival analysis displayed in Supplemental Fig. 2 demonstrates noninferior overall survival among patients with disappearing SRMs compared to others undergoing AS (5-year survival: 100% vs. 73.2%, P = 0.06). 4. Discussion Our study quantifies the rates of disappearing SRMs in a prospective cohort and offers several interesting findings
regarding the management of SRMs on AS. First, 5.9% (22/374) of SRM patients undergoing AS have at least 1 surveillance image showing no lesion. Point prevalence increased from 1.3%−2.9% over 5 years. Furthermore, as follow-up accrues, 5-year cumulative incidence is 13.3%. On average disappearing SRMs had imaging showing no lesion after 2.0 years of follow-up and most commonly after ultrasound surveillance. Fifty percent of the disappearing lesions reappeared on subsequent CT imaging. When assessing statistical correlates of a disappearing SRM,
Table 2 Predictors of disappearing SRM among DISSRM registry Predictors of disappearance vs. entire registry Univariable Variable
OR
Age BMI Male Mass diameter <1 cm Mass diameter <2 cm Incidentaloma Arterial Hounsfield Units Laterality = Left
1.03 0.96 0.84 4.57 1.67 0.75 1.00 1.90
95% CI 0.99 0.91 0.37 1.47 0.74 0.17 0.98 0.80
Multivariable P value
1.07 1.02 1.91 14.19 3.79 3.28 1.02 4.50
0.12 0.19 0.69 0.01 0.22 0.70 0.68 0.14
OR 0.99 0.95 0.18 10.56 − 0.33 1.00 1.17
95% CI 0.90 0.79 0.02 1.11 − 0.02 0.97 0.15
P value 1.09 1.15 1.90 100.32 − 6.49 1.02 9.05
0.91 0.60 0.16 0.04 − 0.47 0.70 0.88
Predictors of disappearance vs. active surveillance masses only Variable
OR
95% CI
Age BMI Male Mass diameter <1 cm Mass diameter <2 cm Incidentaloma Arterial Hounsfield Units Laterality = Left
0.99 0.98 0.91 2.48 1.12 0.73 1.00 1.66
0.95 0.93 0.38 0.79 0.47 0.16 0.98 0.68
1.03 1.03 2.17 7.80 2.70 3.30 1.02 4.06
P value
OR
95% CI
0.55 0.44 0.84 0.12 0.79 0.68 0.95 0.27
0.95 0.91 0.22 2.64 − 0.45 1.00 1.30
0.84 0.74 0.02 0.27 − 0.02 0.98 0.12
95% CI = 95% confidence interval; BMI = body mass index; OR = odds ratio; SRM = small renal mass.
P value 1.06 1.12 2.62 25.88 − 9.62 1.02 14.26
0.35 0.39 0.23 0.41 − 0.61 0.92 0.83
ARTICLE IN PRESS A. Srivastava et al. / Urologic Oncology: Seminars and Original Investigations 00 (2019) 1−6
tumor size <1 cm was the strongest independent predictor. Given, the importance of tumor size and imaging modality in our analysis, we believe that many of these “disappearance” events stem from limitations of ultrasound and the use of multiple imaging modalities. Lastly, patients with disappearing SRMs did not have compromised outcomes and experienced excellent 5-year overall survival. While no prior series quantify rates of spontaneous tumor regression of RCC, several case reports exist in the current literature [1,3−7]. Based on these previous case reports, Wein et al. estimate this incidence of 0.3%−1% [8]. However our results describe a higher incidence, with an absolute proportion of 5.9% and 5-year cumulative incidence of 13.3%. Scheduled imaging surveillance in our AS protocol increases sensitivity when detecting disappearing SRMs. Additionally, regression of RCC metastases mostly constitute old estimates. A review by Jaisezewska et al. finds that 81.3% (48/59) of prior spontaneous RCC regression reports describe regression of metastases as opposed to 8.5% (5/59) describing primary tumor regression. Differently, our study emphasizes variation in imaging detection or possible clinical regression of primary localized tumors among low risk SRMs amenable to AS [3]. Over half of disappearing SRMs reappear on later imaging. In these cases, “disappearance” does not necessarily represent spontaneous tumor regression, but likely spatial and technical limitations of surveillance imaging—particularly renal ultrasound. However, as most tumors on AS are small cT1a lesions, risk of pT3a upstaging or development of metastatic disease is minimal [9,10]. Thus, while ultrasound may have inferior evaluation of the regional lymph nodes and extra renal fat invasion, it safely reduces cost and radiation for those patients on AS. Furthermore, previous studies demonstrate similar renal mass size estimates between ultrasound, CT, and MRI [11]. However, for very small SRMs, which stress the boundaries of imaging resolution, the user-dependent nature of ultrasound may result in more false negative imaging than axial imaging. Also patients with renal medullary necrosis may have obscuring of their renal mass, even on axial imaging delayed contrast phases. Consequently, most disappearing SRMs were encountered on ultrasound and most disappearing masses were <1 cm. Our study may also support that this low risk subset of patients, may actually require less rigorous imaging, than those with larger masses. Interestingly, while not statistically significant, those patients with disappearing masses trended towards having longer follow-up (3.3 vs. 2.5 years, P = 0.08). This may represent patient concerns regarding their mass and clinician efforts to make sure the mass the is recharacterized on subsequent imaging. Therefore, we reassure patients that their SRM may not be “gone” and inform them reappearance is likely on subsequent imaging. Understanding that patients with disappearing SRMs do not experience poor outcomes, we recommend that urologists not follow-up tumor disappearance with immediate reflex imaging. Instead, patients should continue
5
on their original surveillance schedule, sparing them additional radiation and healthcare costs. Finally, our study provides further evidence that AS is safe for SRMs, especially those <1 cm in diameter. 11.4% (4/35) of tumors <1 cm in the DISSRM registry had negative surveillance imaging at least once. Most importantly, disappearing SRMs demonstrate favorable tumor biology and behavior. Not requiring biopsy at enrollment, the DISSRM registry includes patients with malignant and benign tumors. Given the diagnostic limitations of imaging and renal mass biopsy, particularly for SRMs ≤1 cm, this strategy continues to prove safe. For disappearing SRMs, linear growth rates are slow, no patient has developed metastatic disease, and 5-year survival is 100%. SRMs that truly disappear may represent spontaneously resolving benign processes (i.e. infectious or inflammatory) or indolent RCC regressing due to immune surveillance. The improved overall survival reflects good patient health and well-functioning immune systems. A large population-based study examining perioperative infection after nephrectomy hypothesizes that immunologic defenses and prevention of T cell anergy may improve RCC outcomes [12]. Similarly, low risk renal masses may permanently regress due to up-regulation of inflammatory cytokines and cytotoxic T cells. However, we note that heterogenous surveillance imaging, difficult to characterize masses, and limitations in spatial resolution may primarily drive many of these “disappearance" events. Several potential limitations should be considered for the present report. First, the overall proportion of disappearing SRMs remains low, under 6%—but is significantly higher in patients with SRMs <1 cm. Furthermore, as previously described, many of these “disappearing” SRMs occur when using ultrasound and represent limitations in current imaging. With future imaging and followup, these lesions may reappear and depress the incidence we have calculated. Also, sample size may limit conclusions regarding incidence and relative survival of patients with disappearing SRMs. While additional pathological data would be informative, this report demonstrates that very small SRMs, of 1 and 2 cm, had the highest likelihood of “disappearance.” Lastly, we utilized overall survival as a measure of outcomes amongst patients with disappearing masses. Given how few patients truly die from their SRM, it is unlikely that our analysis would uncover a large difference in survival. However, it does allow us to demonstrate that those patients with disappearing SRMs have not been managed unsafely and do not require increased imaging surveillance. This supports recent guidelines for AS as initial management of patients with SRMs ≤2 cm [13]. Our report provides additional evidence regarding the indolent nature of these tumors, questions the ability of current imaging techniques to accurately discriminate very small SRMs, and informs patients and providers that a small, but significant proportion of SRMs may clinically regress.
ARTICLE IN PRESS A. Srivastava et al. / Urologic Oncology: Seminars and Original Investigations 00 (2019) 1−6
6
5. Conclusion AS for SRMs ≤2 cm does not compromise oncologic outcomes for patients with suspected RCC. Approximately 5% of SRMs on AS may have their lesion disappear on surveillance imaging, particularly after ultrasound and in those tumors <1 cm in diameter. Many tumors reappear on subsequent imaging indicating limitations in spatial imaging and potential user dependence. However, patients with disappearing SRMs do not experience compromised outcomes, and a proportion may spontaneously regress. Consequently, disappearing SRMs do not require reflex axial imaging and patients can continue on their original imaging schedule. Supplementary materials Supplementary material associated with this article can be found in the online version at https://doi.org/10.1016/j. urolonc.2019.10.005. References [1] Jawanda GG, Drachenberg D. Spontaneous regression of biopsy proven primary renal cell carcinoma: a case study. Can Urol Assoc 2012;6(5):E203–5. https://doi.org/10.5489/cuaj.11035. [2] Pierorazio PM, Johnson MH, Ball MW, Gorin MA, Trock BJ, Chang P, et al. Five-year analysis of a multi-institutional prospective clinical trial of delayed intervention and surveillance for small renal masses: the DISSRM registry. Eur Urol 2015;68(3):408–15. https://doi.org/ 10.1016/j.eururo.2015.02.001. [3] Janiszewska AD, Poletajew S, Wasiutynski A. Reviews spontaneous regression of renal cell carcinoma. Wspo»czesna Onkol 2013;2(2): 123–7. https://doi.org/10.5114/wo.2013.34613.
[4] Snow RM, Schellhammer PF. Spontaneous regression of metastatic renal cell carcinoma. Urology 1982;20(2):177–81. https://doi.org/ 10.1016/0090-4295(82)90356-9. [5] Oliver RTD, Nethersell ABW, Bottomley JM. Unexplained spontaneous regression and alpha interferon as treatment for metastatic renal carcinoma. Br J Urol 1989;63(2):128–31. https://doi.org/10.1111/ j.1464-410X.1989.tb05147.x. [6] de Riese W, Goldenberg K, Allhoff E, Stief C, Schlick R, Liedke S, et al. Metastatic renal cell carcinoma (RCC): spontaneous regression, long-term survival and late recurrence. Int Urol Nephrol 1991; 23(1):13–25: Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/ 1938215. [7] Srivastava A, Meyer AR, Pierorazio PM, Rowe SP, Allaf ME, Gorin MA. Spontaneous regression of a low-grade renal cell carcinoma with oncocytic features after renal mass biopsy. Clin Genitourin Cancer 2018. https://doi.org/10.1016/j.clgc.2018.07.012. [8] Wein AJ, Kavoussi LR, Novick AC, Partin AW 2012, Peters Ca. Campbell-Walsh Urology. 10th Edition. Campbell-Walsh Urology. https://doi.org/10.1016/B978-1-4160-6911-9.00061-X. [9] Srivastava A, Patel HD, Joice GA, Semerjian A, Gorin MA, Johnson MH, et al. Incidence of T3a up-staging and survival after partial nephrectomy: size-stratified rates and implications for prognosis. Urol Oncol 2018;36(1):7–13. https://doi.org/10.1016/j.urolonc.2017. 09.005. [10] Smaldone MC, Kutikov A, Egleston BL, Canter DJ, Viterbo R, Chen DYT, et al. Small renal masses progressing to metastases under active surveillance. Cancer 2012;118(4):99–1006. https://doi.org/10.1002/ cncr.26369. [11] Mucksavage P, Ramchandani P, Malkowicz SB, Guzzo TJ. Is ultrasound imaging inferior to computed tomography or magnetic resonance imaging in evaluating renal mass zize? Urology 2012;79(1): 28–31. https://doi.org/10.1016/J.UROLOGY.2011.09.036. [12] Tallman JE, Pearce SM, Kuchta K, Helfand BT, Eggener SE. Impact of perioperative infection on cancer specific survival after nephrectomy for renal cell carcinoma. J Urol 2017;198(5):1027–32. https:// doi.org/10.1016/J.JURO.2017.05.070. [13] Campbell S, Uzzo RG, Allaf ME, Bass EB, Cadeddu JA, Chang A, et al. Renal mass and localized renal cancer: AUA guideline. J Urol 2017;198(3):520–9. https://doi.org/10.1016/j.juro.2017.04.100.