Author + information
- Received May 16, 2018
- Revision received September 19, 2018
- Accepted October 18, 2018
- Published online March 25, 2019.
- Takehiro Nakahara, MD, PhDa,b,c,
- Jagat Narula, MD, PhDa,∗∗ (, )@LubbDup@IcahnMountSinai,
- Jan G.P. Tijssen, MD, PhDd,
- Sunil Agarwal, MD, PhDa,
- Mohammed M. Chowdhury, MDe,
- Patrick A. Coughlin, MDe,
- Marc R. Dweck, MD, PhDf,
- James H.F. Rudd, MD, PhDg,
- Masahiro Jinzaki, MD, PhDc,
- John Mulhall, MDh and
- H. William Strauss, MDa,b,∗ (, )@sloan_kettering
- aMount Sinai Heart, Icahn School of Medicine at Mount Sinai, New York, New York
- bMolecular Imaging and Therapy Service, Memorial Sloan Kettering Cancer Center, New York, New York
- cDepartment of Diagnostic Radiology, Keio University School of Medicine, Shinjuku-ku, Tokyo, Japan
- dDepartment of Cardiology, Academic Medical Center–University of Amsterdam, Amsterdam, the Netherlands
- eDepartment of Vascular Surgery, University of Cambridge, Cambridge, United Kingdom
- fBHF Centre for Cardiovascular Science, University of Edinburgh, Edinburgh, United Kingdom
- gDivision of Cardiovascular Medicine, University of Cambridge, Cambridge, United Kingdom
- hSexual and Reproductive Medicine Program, Memorial Sloan Kettering Cancer Center, New York, New York
- ↵∗Address for correspondence:
Dr. H. William Strauss, Molecular Imaging and Therapy Section, Memorial Sloan-Kettering Cancer Center, 1275 York Avenue, New York, New York 10065.
- ↵∗∗Dr. Jagat Narula, Icahn School of Medicine at Mount Sinai, One Guggenheim Pavilion, 1190 Fifth Avenue, Mount Sinai Heart, N-126, Box 1030, New York, New York 10029.
Background Fluorine-18 sodium fluoride (NaF), a bone-seeking radiopharmaceutical used to detect osseous metastases, localizes in regions of microcalcification in atherosclerosis.
Objectives To determine if atherosclerosis of penile arteries plays a role in erectile dysfunction (ED), this study analyzed NaF images in prostate cancer patients.
Methods NaF positron emission tomography–computed tomography bone scans were evaluated in 437 prostate cancer patients (age 66.6 ± 8.7 years). Their urologic histories were reviewed for prevalent ED (diagnosed before the scan date) or incident ED (no ED at first scan, but developed during 1-year follow-up); patients with no ED (neither before the scan nor during follow-up) were included as a control group. A semicircular region of interest was set on the dorsal one-half of the penis (to avoid residual excreted activity in the urethra) on 5 contiguous slices at the base of the penis on positron emission tomography–computed tomography coronal reconstructions, and the average standardized uptake value (SUVmax) was described as NaF uptake.
Results Of 437 patients, 336 (76.9%) had prevalent ED, 60 incident ED (13.7%), and 41 had no ED (9.4%). SUVmax in patients with prevalent (median 1.88; interquartile range [IQR]: 1.67 to 2.16) or incident (median 1.86; IQR: 1.72 to 2.08) ED was significantly higher than no ED (median 1.42; IQR: 1.25 to 1.54) patients (p < 0.001). After adjustment for other risk factors, the odds ratio of prevalent or incident ED was 25.2 (95% confidence interval: 9.5 to 67.0) for every 0.5-U increment in SUVmax with receptor operating characteristic area of 0.91 (95% confidence interval: 0.88 to 0.94).
Conclusions NaF uptake in penile vessels suggests that atherosclerosis is associated with ED in prostate cancer patients. The importance of NaF uptake needs to be tested in noncancer subjects and cause-effect relationship needs to be established.
Erectile dysfunction (ED) shares multiple risk factors with atherosclerosis (1). Previous studies have demonstrated the relationship of ED to increasing age, obesity, smoking, hypertension (HTN), metabolic syndrome, diabetes mellitus (DM), and dyslipidemia (2–4). Endothelial dysfunction, which is commonly linked to all risk factors, is considered to play a major role in ED (5,6), and may be the precursor of early atherosclerotic processes. To determine if atherosclerotic involvement of penile vasculature contributes to ED, we analyzed Fluorine-18 sodium fluoride (NaF) positron emission tomography (PET)–computed tomography (CT) images in prostate cancer patients. NaF, which is a bone-seeking radiopharmaceutical and is used to detect osseous metastases, is also known to localize in the focal areas of active microcalcification in atheromatous plaques (7–10).
We hypothesized that NaF uptake in penile vessels could serve as a marker of vasculogenic ED. During our review of bone scans, we had frequently noted NaF uptake in penile arteries, which led to the current hypothesis. The present study investigates the relationship of penile vascular NaF uptake in patients with and without ED.
The institutional review board approved this retrospective study and waived the necessity for a written informed consent. Prostate cancer patients (n = 442) who had NaF bone scans for detection of osseous metastases were enrolled in this study. The urologic histories were reviewed to determine the presence of ED at the time of initial and 1-year follow-up scans. ED was present when a diagnosis had been established by an experienced urologist or the International Index of Erectile Function (IIEF-5) score was <21 in the absence of any drugs known to cause ED. A total of 5 patients with penile implants were excluded from analysis, reducing the study cohort to 437 patients. Patients were divided into 3 groups: 1) prevalent ED, that is, ED diagnosed before the date of scan; 2) incident ED, that is, no ED at the time of the first scan, but ED developed during 1-year follow-up; and 3) no ED, that is, ED neither at the first date nor during follow-up; this group was treated as the control group. Coronary risk factors on the scan dates were also obtained.
PET-CT protocols and image reconstruction
Approximately 70 min after intravenous injection of 5 to 6 mCi 18F-NaF, whole body PET-CT images were acquired on integrated PET-CT systems (Discovery 690 or 710, GE Healthcare, Milwaukee, Wisconsin). A low-dose CT (120 kVp, 40 to 100 mA based on the body weight, 0.8 s/rotation, and 3.75-mm slice thickness) was performed for attenuation correction and anatomic registration. No iodinated contrast material was administered. PET images were acquired from the vertex to the feet in 3-dimensional mode for 3 min per bed position. PET-CT images were transferred to an offline workstation and reconstructed into coronal, axial, and sagittal planes with dedicated software (Hermes hybrid viewer software, Hermes Medical Solutions, Stockholm, Sweden).
To evaluate NaF uptake in penile arteries at the base of the penis 5 consecutive coronal PET-CT images and the corresponding CT were analyzed. The first slice was at the anterior-inferior margin of the symphysis pubis, and incremented anteriorly by 1 PET slice (with corresponding CT) for each of the remaining 4 slices. The images were evaluated for both calcifications on CT and fluoride PET uptake in the dorsal artery and cavernosal vessels (PET). To exclude residual excreted activity in urine within the penile urethra, a semicircle region of interest (ROI) was set on the upper one-half of the penis (Central Illustration). The average standardized uptake value (SUVmax) (11) of 5 ROIs was defined as the NaF uptake. The comparable slices on CT were evaluated for CT calcification (>130 HU) using an OsiriX workstation (OsiriX version 32 bit, OsiriX Imaging Software, Geneva, Switzerland). This software was also used to measure CT calcification in the common carotid (CCA) and common/internal iliac (CIA/IIA) arteries. For measurement of coronary calcification, we used the 0 to 12 scoring method developed for nongated scans (12,13).
Intraobserver and interobserver variability in NaF measurements was assessed using Bland-Altman analysis and Spearman’s rank correlation coefficients. Data are presented as median (interquartile range [IQR]; i.e., 25th to 75th percentile, or Q1, Q3) or as mean ± 1 SD. Continuous data were compared using Mann-Whitney U test between the 2 groups or Kruskal-Wallis test with post ad-hoc Dwass, Steel, Critchlow-Fligner multiple comparison analysis. Proportions were compared with Fisher exact probability test. We examined differences in SUVmax in various groups including in patients with extra-capsular extension of their prostate cancer or evidence of neuronal involvement of histopathology. The Spearman Rank Correction coefficient test was used for the assessment of linear correlation of 2 parameters. We used multivariable logistic regression model to study the relationship between ED and NaF SUVmax after adjusting for potential confounders. The areas under the receiver-operating characteristic (ROC) curve were used to compare incremental diagnostic utility of NaF for diagnosis of prevalent or incident ED. A 2-sided p < 0.05 was considered statistically significant. Statistical analyses were performed with SAS software, version 9.4. (SAS Institute Inc., Cary, North Carolina).
Of the 437 patients with prostate cancer (age 66.6 ± 8.7 years, body mass index [BMI] 28.9 ± 4.4 kg/m2), 244 patients had hypertension (HTN) (56%), 247 dyslipidemia (57%), 70 type 2 diabetes mellitus (DM) (16%), and 204 reported a smoking history (47%; 26 currently smokers); 52 (12%) patients had a history of coronary artery disease (Table 1). Based on histopathology, ED patients with (n = 165) or without (n = 31) extracapsular extension of prostate cancer and with (n = 173) or without (n = 7) evidence of neuronal involvement were also separately evaluated and compared with the no ED group. As presented in Figure 1, 220 patients had radical prostatectomy prior to the NaF scan (4.1 ± 4.4 years) and 45 patients had radical prostatectomy within 1 year after the (NaF) PET-CT. A total of 145 patients were treated with radiation (25 seed implant and 131 external beam radiotherapy [EBRT]); interval from initial EBRT to NaF scan was 7.7 ± 5.0 years, from salvage EBRT to NaF scan was 4.6 ± 3.7 years, from brachytherapy to NaF scan was 8.1 ± 4.1 years, and the total was 6.2 ± 4.6 years. A total of 139 patients received androgen deprivation therapy (ADT) with or without other radiation therapy and surgical intervention within 6 months. Finally, 92 patients were monitored for follow-up without any active surgical, radiation, or hormonal therapy.
NaF uptake and ED
Penile NaF uptake was commonly observed in the ED patients (Figure 2); the uptake was observed distinctly from the urethral activity in the cavernous and dorsal penile arteries. Former vessels are considered important for cavernous filling and erectile competence. Quantitative penile NaF uptake in prevalent ED (SUVmax 1.88; IQR: 1.67 to 2.16) and incident ED (SUVmax 1.86; IQR: 1.72 to 2.08) was significantly higher than in no ED (SUVmax 1.42; IQR: 1.25 to 1.54) patients (p < 0.001) (Figure 3). Quantification of penile NaF uptake demonstrated excellent interobserver and intraobserver reproducibility (Online Figure 1). The age-adjusted mean difference in SUVmax was higher by 0.53 (IQR: 0.37 to 0.60) among prevalent ED and by 0.56 (IQR: 0.44 to 0.61) among incident ED patients compared with no ED. In our study cohort, IIEF scores were available for 176 patients at baseline and 1-year follow-up IIEF score for 90 patients. NaF uptake was negatively correlated with IIEF score (R = −0.26; p < 0.001; n = 144 prevalent ED, 13 no ED) at baseline and at 1-year follow-up (R = −0.32; p < 0.001; n = 69 prevalent ED, 11 incident ED, 10 no ED patients). On the other hand, although higher, there was no statistically significant difference in CT-verified calcium score in the carotid, internal iliac, common iliac, and penile arteries among prevalent and incident ED compared with no ED. In the coronary arteries (on the ungated scale of 0 to 12 score), the calcium score showed significant difference in 3 groups, it was higher in prevalent ED patients than in no ED patients albeit statistically nonsignificant (p = 0.07) in multiple comparisons (Table 1). Penile NaF uptake was not significantly different in ED patients with (median 1.86 [IQR: 1.64 to 2.10]; n = 165) or without extracapsular extension (median 1.83 [IQR: 1.61 to 2.01]; n = 31), nor in ED patients with (median 183 [IQR: 1.63 to 2.10]; n = 173) or without perineural invasion (median 1.91 [IQR: 1.66 to 2.29]; n = 7).
Although prevalent ED patients were older (67.5 ± 8.3 years) than no ED (64.0 ± 9.4 years) or incident ED (63.5 ± 9.6 years), there was only a modest correlation between age and SUVmax (R = 0.22; p < 0.001). The BMI was not significantly different between no ED (28.1 ± 3.8 kg/m2), incident ED (29.4 ± 5.0 kg/m2), and prevalent ED (29.0 ± 4.3 kg/m2) groups but showed a significant correlation with the SUVmax (R = 0.37; p < 0.001). The estimated glomerular filtration rate (Modification of Diet in Renal Disease Study) was not significantly different between the no ED (84.4 ± 23.9 ml/min/1.73 m2), incident ED (85.2 ± 19.4 ml/min/1.73 m2), and prevalent ED (82.9 ± 20.5 ml/min/1.73 m2) and showed no correlation with the SUVmax (R = −0.08; p =0.12) (Table 1). SUVmax was significantly higher in patients with hypertension, diabetes, and prior radiation therapy. SUVmax was not different in patients with a history of smoking, coronary artery disease (CAD), and/or radical prostatectomy. Dyslipidemia did not affect SUVmax and, in the subgroup with dyslipidemia, statin did not affect SUVmax (Online Table 1).
After adjustment for age, BMI, eGFR, hypertension, hyperlipidemia, diabetes mellitus, smoking status, prevalent CAD, prior radiation therapy, prior prostate surgery, and ADT, the odds ratio of any ED (prevalent or incident vs. no ED) was 25.2 (95% CI: 9.5 to 67.0) for each 0.5-U increment in SUVmax (Online Table 2). Only a negligible change occurred after adding all other clinical variables, including CT calcification in the coronary, carotid, common/internal iliac arteries, and penile arteries to this model. The strong relationship between SUVmax and ED was also reflected in area under ROC of 0.91 (95% CI: 0.88 to 0.94) in a model with SUVmax alone as predictor (Figure 4).
Effect of therapeutic strategy on NaF uptake
Because ED may also result as a consequence of the treatment, such as nerve injury, we reviewed the results separately in groups of patients treated by surgical, radiation, or hormonal protocols (14). In all subgroups, SUVmax was higher for the prevalent and incident ED including only radiation (median 2.01 [IQR: 1.83 to 2.40]; n = 36), only prostatectomy (median 1.81 [IQR: 1.59 to 2.10]; n = 111), only androgen deprivation therapy (median 1.94 [IQR: 1.69 to 2.16]; n = 40) and those only under surveillance without any surgical, radiation, or hormonal intervention (median 1.84 [IQR: 1.63 to 2.16]; n = 79) compared with the no ED patients (p < 0.05) (Figure 3). Within 1 year after the (NaF) PET-CT, 45 patients underwent radical prostatectomy, 46 patients received radiation therapy, and 153 patients were treated with ADT. However, these therapies within 1 year after the NaF imaging did not influence incident ED. Medications, especially beta-blockers and anxiolytics, and antidepressant agents including selective serotonin inhibitors, which could potentially cause ED, did not influence the prevalent or incident ED (p = NS) (Online Table 3).
NaF imaging in erectile dysfunction
This study showed that penile NaF uptake was associated with the presence of ED and the likelihood of future ED. It is possible that penile NaF uptake is an indicator of penile vascular pathology and, hence, vasculogenic ED. Using SUVmax of 1.56 as the cutoff value based on the ROC analysis, NaF uptake showed sensitivity of 85%, specificity 80%, positive predictive value of 98%, and negative predictive value 35%, with diagnostic accuracy of 84%. The low negative predictive value could represent the patients with nonvasculogenic ED, such as from psychological, neurological, or hormonal causes (1–3,14). A subanalysis of MESA (Multiethnic Study of Atherosclerosis) showed coronary calcium score to be an important predictor of endothelial and erectile dysfunction (15). Although our study population was smaller (n = 437, CAD = 52) compared with 1,862 men (age 45 to 84 years, free of CAD) in the MESA study, our results of calcium scores were compatible with the MESA study. The ROC analysis and odds ratio showed that NaF ED was superior to coronary calcium score to predict ED.
Pathogenetic basis of NaF uptake in ED
NaF has been used for >50 years (16) as a tracer to detect osseous metastasis. It has also been proposed that NaF uptake could identify the active process of microcalcification in atheromatous plaques (7,17,18). The initial step of microcalcification in atheroma involves extracellular vesicles, including matrix vesicles and apoptotic bodies in the necrotic core (19), which serve as nucleation sites for calcium phosphate deposition. As microcalcification increases, it coalesces into large masses or triggers a calcification cascade (as seen in bone formation) and results in macrocalcification (10,20).
There are no reports describing the use of NaF to evaluate possible atherosclerosis in erectile dysfunction. The artery size hypothesis suggests that a relatively small decrease in penile artery diameter, such as due to plaque, could cause ED (21). In an autopsy study of 31 subjects, the prevalence of atherosclerotic lesions in penile arteries (12.9%) was lower than that of the coronary arteries (87.1%) and internal iliac arteries (77.4%) (22). AHA type Vb atherosclerosis was observed in penile arteries, wherein the mineral deposition is expected to have replaced the contents of the atheromatous necrotic core. From these observations, it is conceivable that NaF should localize in early microcalcific foci in penile vessels. The fluoride uptake provides information that differs from late calcification seen on CT (10).
Due to the limited spatial resolution (3 to 5 mm) of PET imaging, it is not certain if the penile NaF uptake is localized to cavernous arteries. Since there was no significant correlation between iliac calcification and penile fluoride uptake, it is also possible that calcification could be associated with the venous compartment or smooth muscle cell damage leading to possible venous insufficiency and ED.
Inflammation as the basis of vascular involvement and NaF uptake
Inflammation, which is the obligatory component of atherosclerosis, has been proposed as the precursor of microcalcification and NaF uptake (10). Although nerve injury from prostatectomy has been widely implicated as an etiology of ED, Mulhall et al. (23) have emphasized that ED is associated with arterial insufficiency and venous leakage, especially when the cavernous nerve remained unaffected during radical prostatectomy. The result of the current study with high prevalence of NaF uptake in the penile vasculature in ED patients is consistent with their data. Further, the relationship between ED and external beam radiation was reported in the 1970s, with impotence occurring several months to several years after therapy (24,25), possibly an inflammatory consequence of irradiation. Our study demonstrated that patients who received radiation therapy >3 years before NaF scan showed a high odds ratio in univariate analysis (odds ratio: 12.3 [95% CI: 1.7 to 90.5]), but did not achieve statistical significance in multivariable analysis (Online Table 2). It is possible that vascular risk factors act synergistically with radiation to result in ED (25). Animal studies have demonstrated that radiation-induced inflammation could accelerate development of atherosclerotic lesions in ApoE−/− mice (26,27).
To evaluate if inflammation was a necessary accompaniment of NaF uptake in ED patients, we reviewed a subset of 63 patients who had undergone both (FDG) PET-CT and (NaF) PET-CT scans within 6 months of each other (Online Figure 2); FDG uptake is an established marker of vascular inflammation. There was, however, no significant difference of penile FDG uptake in the no ED (median 1.52 [IQR: 1.09 to 2.29]; n = 7), incident ED (median 1.64 [IQR: 1.33 to 2.07]; n = 11), or prevalent ED (median 1.95 [IQR: 1.63 to 2.40]; n = 45) groups (p = 0.08). Also, there was no significant correlation between NaF uptake and FDG uptake (r = 0.15; p = 0.24). However, SUVmax of FDG was significantly higher in patients treated with irradiation (median 2.29 [IQR: 1.85 to 2.42]; n = 19) compared with those without radiation therapy (median 1.77 [IQR: 1.51 to 2.05]; n = 44; p < 0.05); NaF imaging also demonstrated a trend (albeit statistically insignificantly) toward higher uptake in the radiation group supporting an inflammatory basis. In all other patients, ED correlated with increased penile fluoride uptake, but not with FDG uptake, suggesting that the calcification may not necessarily be associated with the inflammatory stage of the atherosclerotic process. Medial and intimal calcification is commonly seen in peripheral artery disease regardless of inflammation (28) and may also contribute to penile NaF uptake.
Management strategy, ED and NaF uptake
In our series of unselected prostatic malignancy patients, variable degrees of prevalent ED were reported in 77% of patients, which might seem high. A plausible approach to evaluate the true prevalence of ED in the prostate cancer patients is to compare no ED patients with incident ED patients, because all patients in these 2 groups should not have ED at the time of initial scan, the incident ED patients developed ED during follow-up. Whereas 41 patients belonged to no ED group, 60 patients had incident ED; approximately 60% of patients with no evidence of ED at baseline developed ED within the ensuing year. The impotency rates have been reported to range from 54% to 90% in the 12 months of follow-up after radical prostatectomy with or without robot assistance (29), and approximately 50% of patients at 5 years after radiation therapy (30). The prevalence of ED in a normal population (mean age 62.3 years) has been reported as 52% (31). Considering our patients were older than 66 years, the prevalence of ED in our study is not an overestimate. In the subgroup of no ED versus incident ED (n = 101) patients, ROC curve analysis showed a high diagnostic accuracy of NaF SUVmax for incident ED, with the area under the curve of 0.91 (95% CI: 0.86 to 0.97) (Online Figure 3).
Although the results are intriguing and provocative, there are several limitations of this study. It is a retrospective study of prostate cancer patients from a single tertiary care referral center, and the inherent bias of inclusion of only severe and complicated cases cannot be excluded. In addition, the number of no ED patients is small. However; considering published rates of ED in patients with prostate cancer as discussed in the previous text, it appears that we have sampled a representative group of patients with a high prevalence of ED. Regardless, the lower number of no ED patients reduces the confidence in our findings. Unfortunately, most of our patients did not have lipid profiles, so we could not calculate Framingham risk scores to evaluate relationship between risk factors and possible penile artery atherosclerosis. Therefore, this study still leaves open the possibility that the NaF uptake is a consequence of or a co-occurrence from ED-causing mechanisms (i.e., risk factors and endothelial dysfunction) rather than a causative etiology of ED. Finally, even though there is a higher likelihood of active penile arterial calcification in ED, the study does not prove causality nor reveals clues that may help manage prevalent ED or prevent occurrence of incident ED.
Fluoride uptake in penile vessels is significantly higher in patients with prevalent or incident ED. NaF uptake occurs both in patients undergoing surgical or radiation therapy. The study only demonstrates a correlation, and causative association needs to be established. Therefore, future studies would examine the role of penile vascular fluoride uptake as a contributor to ED especially in a general patient population.
COMPETENCY IN MEDICAL KNOWLEDGE: 18F-sodium fluoride uptake is strongly related to incident and prevalent ED in prostate cancer patients.
TRANSLATIONAL OUTLOOK: Future studies should examine the association of penile vascular fluoride uptake with ED in relation to age, comorbidities, and therapeutic strategies in the general population.
The authors thank Josef J. Fox, MD, Mr. J. Kalaigian, and Mr. C. Qing for their technical support.
Dr. Nakahara was supported by the SNMMI Wagner-Torizuka Fellowship and Uehara Memorial Foundation Fellowship. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose. Nikhil Joshi, MD, PhD, served as Guest Editor for this paper.
Listen to this manuscript's audio summary by Editor-in-Chief Dr. Valentin Fuster on JACC.org.
- Abbreviations and Acronyms
- coronary artery disease
- external beam radiotherapy
- erectile dysfunction
- Fluorine-18 sodium fluoride
- positron emission tomography–computed tomography
- receptor operating characteristic
- region of interest
- standardized uptake value
- Received May 16, 2018.
- Revision received September 19, 2018.
- Accepted October 18, 2018.
- Fung M.M.,
- Bettencourt R.,
- Barrett-Connor E.
- Keenan H.A.
- Kaiser D.R.,
- Billups K.,
- Mason C.,
- Wetterling R.,
- Lundberg J.L.,
- Bank A.J.
- Derlin T.,
- Richter U.,
- Bannas P.,
- et al.
- Fiz F.,
- Morbelli S.,
- Piccardo A.,
- et al.
- Irkle A.,
- Vesey A.T.,
- Lewis D.Y.,
- et al.
- Nakahara T.,
- Dweck M.R.,
- Narula N.,
- Pisapia D.,
- Narula J.,
- Strauss H.W.
- Weber W.A.
- Yankelevitz D.F.,
- Henschke C.I.,
- Yip R.,
- et al.
- Feldman D.I.,
- Cainzos-Achirica M.,
- Billups K.L.,
- et al.
- Dweck M.R.,
- Chow M.W.,
- Joshi N.V.,
- et al.
- New S.E.,
- Goettsch C.,
- Aikawa M.,
- et al.
- Nakahara T.,
- Strauss H.W.
- O'Neill W.C.,
- Han K.H.,
- Schneider T.M.,
- Hennigar R.A.
- Gaither T.W.,
- Awad M.A.,
- Osterberg E.C.,
- et al.