Predicting ¹⁸F-DCFPyL-PET/CT Scan Positivity in Prostate Cancer Patients with Biochemical Recurrence

Abstract

Background:

To analyze variables that can predict the positivity of ¹⁸F-DCFPyL-positron emission tomography/computed tomography (PET/CT) and extent of disease in patients with biochemically recurrent (BCR) prostate cancer after primary local therapy with either radical prostatectomy or radiation therapy.

Materials and Methods:

This is a retrospective analysis of a prospective single institutional review board-approved study. We included 199 patients with biochemical recurrence and negative conventional imaging after primary local therapies (radical prostatectomy n=127, radiation therapy n=72). All patients underwent ¹⁸F-DCFPyL-PET/CT. Univariate and multivariate logistic regression analyses were used to determine predictors of a positive scan for both cohort of patients. Regression-based coefficients were used to develop nomograms predicting scan positivity and extra-pelvic disease. Decision curve analysis (DCA) was implemented to quantify nomogram’s clinical benefit.

Results:

Of the 127 (63%) post-radical prostatectomy patients, 91 patients had positive scans – 61 of those with intrapelvic lesions and 30 with extra-pelvic lesions (i.e., retroperitoneal or distant nodes and/or bone/organ lesions). Of the 72 post-radiation therapy patients, 65 patients had positive scans – 39 of them had intrapelvic lesions and 26 extra-pelvic lesions. In the radical prostatectomy cohort, multivariate regression analysis revealed original International Society of Urological Pathology category, prostate-specific antigen (PSA), prostate-specific antigen doubling time (PSAdt), and time from BCR (mo) to scan were predictors for scan positivity and presence of extra-pelvic disease, with an area under the curve of 80% and 78%, respectively. Positive vs negative tumor margin after radical prostatectomy was not related to scan positivity or to the presence of positive extra-pelvic foci. In the radiation therapy cohort, multivariate regression analysis revealed that PSA, PSAdt, and time to BCR (mo) were predictors of extra-pelvic disease, with area under the curve of 82%. Because only 7 patients in the radiation therapy cohort had negative scans, a prediction model for scan positivity could not be analyzed and only the presence of extra-pelvic disease was evaluated.

Conclusions:

PSA and PSAdt are consistently significant predictors of ¹⁸F-DCFPyL PET/CT positivity and extra-pelvic disease in BCR prostate cancer patients. Stratifying the patient population into primary local treatment group enables the use of other variables as predictors, such as time since BCR. This nomogram may guide selection of the most suitable candidates for ¹⁸F-DCFPyL-PET/CT imaging.

Background

After primary local treatment of prostate cancer with either radiation therapy (RT) or radical prostatectomy (RP), 20–50% of patients experience biochemical recurrence (BCR) within 10 years of initial treatment (1–3). For patients who are surgically treated, BCR is defined as rising prostate-specific antigen (PSA) value ≥0.2 ng/mL after RP with a second confirmatory elevated level (4). For those treated with RT as primary therapy, BCR is defined by the Phoenix criteria as any PSA increase >2 ng/ml higher than the PSA nadir (5). Although rising PSA can predict recurrent disease, it cannot localize the recurrence; therefore, sensitive imaging modalities with high diagnostic accuracy are required to distinguish among local relapse, oligometastatic disease (defined as 1–5 lesions in most studies), and extensive disease. The management of these patients includes salvage therapies based on the risk of either isolated pelvic relapse or oligometastatic disease (defined as 1–5 lesions in most studies) with potentially more limited metastatic capacity, less aggressive behavior and better outcomes; versus systemic therapies for those with extensive metastatic disease (6). Thus, identifying the extent of disease spread with imaging is crucial for therapeutic decision-making and mapping potential treatment fields in these patients.

The emergence and increasing availability of prostate specific membrane antigen (PSMA)-positron emission tomography/computed tomography (PET/CT) imaging has shown high sensitivity and specificity for tumor detection in the BCR setting (7,8). ¹⁸F-DCFPyL is a PSMA-targeted radiotracer that was approved by the Food and Drug Administration in 2021 for commercial use in patients with suspected prostate cancer (PCa) metastasis or BCR due to elevated serum PSA (9).

Although guidelines suggest that PSMA-PET/CT is indicated in men with BCR (10), this use-policy may not be the most cost-effective as many men with BCR will have no findings on the scan (11) and the radionuclide itself may cost over $3,000 in the United States. Therefore, a nomogram to predict the probability of having a positive scan could guide clinicians in selecting the most appropriate candidates for PSMA PET/CT imaging.

Prediction models for PSMA PET/CT positivity have been previously published, each analyzing different sets of predictive variables including age, PSA at the time of the scan, PSA doubling time (PSAdt), Gleason score, initial tumor (T) stage, initial nodal (N) stage, initial grade group, salvage radiotherapy after RP, time from treatment to scan, primary therapy (surgery vs radiotherapy), androgen deprivation therapy (ADT) at the time of scan, and International Society of Urological Pathology (ISUP) category (11–16). In addition to PSMA PET/CT scan positivity, prediction models for the presence of extra-pelvic disease have also been studied (16,17). Understanding the extent of the disease (intra-pelvic vs extra-pelvic) is critical in predicting disease prognosis and planning subsequent salvage vs. systemic therapies for these patients. The current literature lacks sufficient data on the use of a predictive model for predicting the presence of extra-pelvic disease on ¹⁸F-DCFPyL-PET/CT, which we believe could be used to better tailor management plans specific to each patient. Additionally, our study stratifies patients by their primary treatment (surgery vs radiation) for comparing the predictive variables for each cohort, since these are different patient populations with very different clinical parameters that need to be assessed separately. In this study, we aim to develop predictive models for predicting the probability of positive ¹⁸F-DCFPyL-PET/CT and of positive extra-pelvic foci in patients with prostate cancer who experience BCR.

Materials and Methods

We included 199 consecutive patients (mean age: 67 years; range: 48–85 years) who signed written informed consent to be enrolled (Trial registration: ClinicalTrials.gov, NCT03181867, registered 6/8/2017, https://clinicaltrials.gov/ct2/show/NCT03181867)(18). All patients had BCR after primary therapy with either RP or RT (median PSA 2.0±5.85 ng/mL) and underwent ¹⁸F-DCFPyL-PET/CT. All patients had negative conventional imaging (CT and bone scan) within 3 months of accrual. All patients had Eastern Cooperative Oncology Group Performance scores of 0 to 2 and histologically confirmed adenocarcinoma of the prostate. Exclusion criteria included current ADT, inability to tolerate a PET/CT scan, or creatinine levels greater than 2 times the upper limit of normal.

All patients received IV injection of ¹⁸F-DCFPyL, mean injected activity of 296±33.3 MBq [8.0±0.9 mCi]. A whole body ¹⁸F-DCFPyL PET/CT was obtained after a 2-hour uptake period (3 min/bed position), using a 3D time of flight (TOF) GE Discovery MI DR scanner, with a 20-cm coronal and a 70-cm axial field of view. Image reconstruction utilized an AC 3D Iterative MLEM algorithm using 29 subsets, 3 iterations, TOF, point spread function regularization parameter 6.0, and a Gaussian post-filter with 4.1-cm kernel. A low-dose non-contrast CT (120 kV, 60 mAs) was acquired with each PET scan for attenuation correction and anatomical co-registration purposes.

Two board-certified nuclear medicine physicians independently reviewed the images, resolving disagreements by consensus. ¹⁸F-DCFPyL-PET/CTs were assessed using MIM (version 7.0.2, MIM Software Inc., Cleveland, OH).

¹⁸F-DCFPyL-PET/CT scans were classified as negative, positive for intra-pelvic disease or positive for extra-pelvic disease and were correlated with clinical variables via a proportional odds model. A scan was considered positive when there were clear foci of abnormal ¹⁸F-DCFPyL uptake above the surrounding background (19), not associated with physiological uptake or other known pitfalls (20). Foci with subtle or mild uptake not definitive for disease were labeled indeterminate and considered negative to avoid confusing results. N, based on PSMA RADS reporting criteria system (21). Indeterminate lesions were based on PET uptake, independent of lesion size. PET positive disease location was classified as intrapelvic (i.e., recurrence was seen in the prostatectomy bed/prostate and/or in pelvic nodes) vs. extra-pelvic (i.e., recurrence was seen in retroperitoneal/distant nodes and/or bone/organ lesions).

Clinical variables of interest included age, ISUP category, PSA, PSA doubling time (PSAdt) (mo), time since last treatment, time since BCR (mo), type of treatment (RP vs RT), T stage, N stage, and tumor margin. To improve the fit of the models to the observed data, PSA and time since BCR were transformed using the base 2 logarithm. Clinical variables which were significant in the univariable analysis were included in the multivariable analysis and variable selection procedure based on the Akaike information criterion was used to determine the final independent predictors for RP cohort and RT cohort, separately. For the RT cohort, as all except seven participants had positive scans, the scan imaging result was dichotomized into extra-pelvic disease or else logistic regression model was used to fit the data and to determine the independent predictors of extra-pelvic disease. The lrm function in the rms R package was used to fit proportional odds ratio models in the radical prostatectomy therapy cohort and glm function with the logit link was used to fit logistic regression models in the radiation therapy cohort. The multivariable model consisting of variables significant in the univariable analysis was selected as the initial model and the backward variable selection procedure was implemented where a variable was eliminated if dropping it from the multivariable model would yield the minimal AIC value. The nomograms developed from the multivariable models were created by the nomogram function in the rms package and were used to predict the likelihoods of scan positivity and extra-pelvic disease. The diagnostic accuracy of the nomogram was measured by area under the curve (AUC) of the receiver operating characteristic curve cross-validated by the bias-corrected bootstrap resampling procedure with 2000 bootstrap samples. PSA and PSAdt values were compared using two sample t-tests. The 95% confidence limits were taken from the 2.5th and 97.5% percentiles of the bootstrap resampling distribution.

Moreover, we implemented the decision curve analysis (DCA) in order to quantify the nomogram’s clinical outcome on clinical practice (22). DCA investigates the theoretical relationship between the threshold probability of positive PSMA PET/CT and the relative value of false-positive and false-negative findings to assess the net benefit of the predictive multivariable model. The clinical utility of the nomograms was measured by net benefit, net reduction in the number of false positives (22), proportion of avoided scans and proportion of missed positive scans. All tests were two-sided and p-value<0.05 was considered statistically significant.

Results

We evaluated 199 patients (mean age: 67 years; range: 48–85 y) who experienced BCR after receiving primary local treatment with either RP or RT: 127 (63.8%) patients were treated with surgery and 72 (36.2%) patients were treated with RT (table 1). None of the patients were receiving androgen deprivation therapy at the time of the scan. Overall, 156 patients (78.4%) had a positive ¹⁸F-DCFPyL PET/CT scan, 91 patients in the post-surgical group and 65 patients in the post-RT group.

Table 1.

¹⁸F-DCFPyL-PET/CT scan findings of all patients stratified by their primary therapy.

Primary Therapy	Positive scan (%)	Negative scan (%)	Intrapelvic disease (%)	Extra-pelvic disease (%)	Total (%)
Radical Prostatectomy (%)	91 (45.7)	36 (18.1)	61 (30.7)	30 (15.1)	127 (63.8)
Radiation Therapy (%)	65 (32.7)	7 (3.5)	39 (19.6)	26 (13.1)	72 (36.2)
Total (%)	156 (78.4)	43 (21.6)	100 (50.3)	56 (28.1)	199 (100)

For the RP cohort (table 2), the mean PSA at the time of scan was 2.49 ± 4.16 ng/ml and the mean PSAdt was 9.56 ± 10.28 mo. Within this cohort, those with a positive ¹⁸F-DCFPyL-PET/CT had a higher mean PSA than those with a negative scan (3.02 ± 4.61 ng/ml vs. 1.14 ± 2.28 ng/ml; p=0.021). Likewise, those with extra-pelvic disease on the scan had a higher mean PSA than the patients with intrapelvic disease (5.59 ± 7.04 ng/ml vs. 1.76 ± 1.72 ng/ml, p<0.01). Post-surgical patients with negative scans also had a longer PSAdt of 12.44 ± 14.53 mo, whereas those with positive scans had a PSAdt of 8.41 ± 7.79 mo (p=0.046). The mean PSAdt of those with intrapelvic disease and extra-pelvic disease were 9.45 ± 8.83 mo and 6.33 ± 4.59 mo, respectively (p=0.073).

Table 2.

PSA and PSA doubling time values of each cohort and subgroups within.

	PSA (ng/ml) (mean ± std)	p-value	PSA doubling time (months) (mean ± std)	p-value
RP	2.49 ± 4.16		9.56 ± 10.28
Positive scan	3.02 ± 4.61	0.021	8.41 ± 7.79	0.046
Negative scan	1.14 ± 2.28		12.44 ± 14.53
Intrapelvic disease	1.76 ± 1.72	0.0001	9.45 ± 8.83	0.073
Extra-pelvic disease	5.59 ± 7.04		6.33 ± 4.59
RT	6.74 ± 7.28		14.43 ± 14.77
Positive scan	7.12 ± 7.53	0.174	13.85 ± 13.75	0.330
Negative scan	3.18 ± 2.56		19.63 ± 22.81
Intrapelvic disease	5.67 ± 5.19	0.057	18.84 ± 15.38	0.0002
Extra-pelvic disease	9.30 ± 9.79		6.27 ± 4.94

Of the 91 scan positive patients in the RP group, 61 (30.7%) patients had intrapelvic disease and 30 (15.1%) had extra-pelvic disease. Univariate regression analysis showed that ISUP category of ≥2, shorter PSAdt, higher initial T stage (≥3a), PSA, more months from BCR to scan were significant predictors for positive scan (table 3). Age, positive vs negative tumor margins after prostatectomy, and N stage were not associated with scan positivity or extra-pelvic disease (p=0.092, p=0.859 and p=0.143, respectively). Multivariate regression analysis revealed ISUP category, PSAdt, PSA, and time since BCR to scan as predictors for scan positivity and for the presence of positive extra-pelvic foci, with an AUC of 80% and 78%, respectively (table 3, figure 1). Finally, the decision curve analysis (DCA) showed the best net clinical benefit when the threshold probability was ≥45 % with a net benefit 54.8% in the ¹⁸F-DCFPyL-PET/CT nomogram, i.e. the net benefit of the ¹⁸F-DCFPyL-PET/CT nomogram was equivalent to performing 54 scans per 100 men without negative scans. A 45% risk threshold was chosen so we can avoid more scans while maintaining the similar risk of missing positive scans (table 4). The net reduction in the number of false-positives with the ¹⁸F-DCFPyL-PET/CT nomogram, compared to performing a scan in all patients was equivalent to performing 8 fewer unnecessary scans per 100 men, without increasing the number of positive scans being missed (figure 2 and table 4).

Table 3.

Univariable and multivariable logistic regression analysis of a positive PSMA PET/CT in the radical prostatectomy therapy cohort

Regression coefficients and odds-ratios (ORs) of Radical Prostatectomy Cohort
	Univariable Analysis		Multivariable Analysis
	OR (95% CI)	p-value	coefficient	OR (95% CI)	p-value
Intercept: scan positive vs scan negative	NA	NA	−0.225	NA	0.797
Intercept: positive extra-pelvic vs else	NA	NA	−3.214	NA	0.001
Age	1.042 (0.993 – 1.094)	0.092
Original ISUP Category
ISUP < 2	1.0 (reference)		0 (reference)	1.0 (reference)
ISUP ≥ 2	4.802 (1.488 – 15.493)	0.009	1.404	4.072 (0.871 – 19.03)	0.074
PSAdt	0.958 (0.924 – 0.994)	0.021	−0.06	0.942 (0.895 – 0.992)	0.023
Surgical margin:
Negative	1.0 (reference)
Positive	1.066 (0.527 – 2.157)	0.859
Initial T stage:
< T3a	1.0 (reference)
≥ T3a	2.74 (1.381 – 5.434)	0.004
Initial N stage:
0	1.0 (reference)
1	2.251 (0.76 – 6.665)	0.143
Time to BCR	0.995 (0.927 – 1.068)	0.89
log2 (PSA)	2.036 (1.587 – 2.611)	<0.001	0.705	2.03 (1.52 – 2.70)	<0.001
log2 (time from BCR)	1.333 (1.105 – 1.609)	0.003	0.225	1.254 (0.97 – 1.62)	0.087
Scan positivity model AUC			80%
Presence of extra-pelvic disease model AUC			78%

Figure 1.

Nomogram predicting the probability of ¹⁸F-DCFPyL-PET/CT positivity and extra-pelvic disease in BCR patients after radical prostatectomy.

Table 4.

Different net benefit according to the different risk threshold for the prostatectomy cohort. The optimum risk threshold was ≥45%. At the 45% risk threshold, the net benefit was 48.5% in the scan-all model and 54.8% in the PSMA nomogram, and net reduction in the number of false-positives was 0% in the scan-all model and 7.8% in the PSMA nomogram. The net benefit of the PSMA nomogram was equivalent to performing 54 scans per 100 men without negative scans. The net reduction in the number of false positives with the PSMA nomogram, compared to performing a scan in all patients was equivalent to performing 8 fewer unnecessary scans per 100 men, with no increase in the number of positive scans being missed. At the 45% risk threshold, per 1,000 men 133 scans could have been avoided while 25 (3.5%) of positive scans would have been missed. We chose the 45% risk threshold so we can avoid more scans while maintaining the similar risk of missing positive scans. Specifically, there was only 1.2% difference in missing positive scans between risk threshold at 35% or less vs 45%, whereas compared to the risk threshold at 35% or less, 4–7% scans can be avoided at the risk threshold 45%.

	Net Benefit (%)	Net benefit (%)	Net reduction in false positives (%)
Risk Threshold	Scan all men	PSMA nomogram	PSMA nomogram
≥30%	59.5	60	1.1
≥35%	56.4	58.8	4.4
≥40%	52.8	56.4	5.4
≥45%	48.5	54.8	7.8
≥50%	43.3	54.2	10.8
Scan all men	Performed	Avoided	Scan positive	Missed (%)
	1000	0	717	0 (0)
Risk by PSMA nomogram
≥30%	933	67	700	17 (2.3)
≥35%	908	92	700	17 (2.3)
≥40%	883	117	692	25 (3.5)
≥45%	867	133	692	25 (3.5)
≥50%	792	208	667	50 (7)

Figure 2.

Decision curve analysis of clinical utility of ¹⁸F-DCFPyL-PET/CT scan positivity for PSMA nomogram model, Scan-all model, and Scan-none model. Net reduction in false positives plotted for PSMA nomogram model in comparison to scanning all patients.

For the RT cohort (table 2), the mean PSA at the time of scan was 6.74 ± 7.28ng/ml and the mean PSAdt was 14.43 ± 14.77mo. Of this cohort, the mean PSA of those with positive and negative ¹⁸F-DCFPyL-PET/CT were 7.12 ± 7.53 ng/ml and 3.18 ± 2.56 ng/ml, respectively (p=0.174). Likewise, the mean PSA of post-RT patients with intrapelvic disease and extra-pelvic disease were 5.67 ± 5.19 ng/ml and 9.30 ± 9.79 ng/ml, respectively (p=0.057). The mean PSAdt of post-RT patients with positive scans was shorter than that of patients with negative scans (13.85 ± 13.75 mo vs. 19.63 ± 22.81 mo (p=0.330)). The PSAdt for post-RT patients with intrapelvic disease was longer than patients with extra-pelvic disease (18.84 ± 15.38mo and 6.27 ± 4.94 mo, respectively; p<0.01).

Because only 7 out of 72 patients in the RT cohort had negative scan results, the prediction model for scan positivity could not be analyzed and only the presence of extra-pelvic disease was assessed for this cohort (table 5). Out of 65 patients with a positive scan in the RT group, 39 (19.6%) had intrapelvic foci and 26 (13.1%) had extra-pelvic lesions. Univariate analysis revealed higher PSA, shorter PSAdt, and shorter time to BCR to be significant predictors of positive extra-pelvic foci on scan, as seen in all cohorts. Unlike the RP cohort, the initial T stage and time since BCR to scan were not significant predictors in the RT cohort. From a multivariate regression analysis, PSA, PSAdt, and time to BCR were predictors for presence of extra-pelvic lesions, with AUC of 82% (table 5, figure 5).

Table 5.

Univariable and multivariable logistic regression analysis for the presence of extra-pelvic disease in the radiation therapy cohort. The N-stage variable was not included in the analysis due to unreliable statistical results from small positive N-stage sample size.

Regression coefficients and ORs of Radiation Therapy Cohort
	Univariable Analysis		Multivariable Analysis
	OR (95% CI)	p-value	Coefficient	OR (95% CI)	p-valu e
Intercept	NA	NA	1.625	NA	0.113
Age	0.988 (0.928 – 1.05)	0.691
Original ISUP Category:
ISUP < 2	1.0 (reference)
ISUP ≥ 2	2.7 (0.863 – 8.451)	0.088
PSA_DT	0.835 (0.752 – 0.928)	0.001	−0.249	0.779 (0.67 – 0.907)	0.001
Initial T stage:
< T3a	1.0 (reference)
≥ T3a	3.583 (0.779 – 16.491)	0.101
Time to BCR	0.625 (0.409 – 0.953)	0.029	−0.648	0.523 (0.273 – 1.003)0.051	0.051
log2 (PSA)	1.498 (1.003 – 2.24)	0.049	0.707	2.028 (1.127 – 3.648)0.018	0.018
log2 (time from BCR)	0.935 (0.704 – 1.243)	0.644
Model AUC			82%

Figure 5.

¹⁸F-DCFPyL-PET/CT imaging of a 73-year-old patient with NCCN high-risk (ISUP GG4) prostate cancer who underwent treatment with neoadjuvant androgen deprivation therapy (ADT) and enzalutamide for 6 months followed by radical prostatectomy with pelvic lymph node dissection. Surgical pathology revealed pT3b N1 R0 disease with extra-prostatic extension, seminal vesicle invasion excised to negative margins; Three of 29 lymph nodes were positive. He then underwent adjuvant radiotherapy to the prostate bed to a dose of 68Gy with 12 months of ADT. Thereafter he experienced biochemical failure and underwent an ¹⁸F-DCFPyL-PET/CT at a PSA of 1.4 ng/mL which revealed two sub-centimeter DCFPyL-avid retroperitoneal lymph nodes consistent with nodal metastases.

Discussion

In this study, we identified parameters associated with positive ¹⁸F-DCFPyL-PET/CT scans after RP or RT. We analyzed data for cohorts divided by their primary treatments: RP (figures 4, 5) and RT (figure 3). Multivariable regression analyses of the RP cohort revealed the following parameters to be significant predictors of a positive scan with or without extra-pelvic disease: ISUP category, PSAdt, PSA, and time from BCR. For the RT cohort, PSA, PSAdt, and fewer time to BCR were predictors for the presence of extra-pelvic disease on ¹⁸F-DCFPyL-PET/CT. PSA and PSAdt have been previously reported to predict for positive scans, mainly using ⁶⁸Ga-PSMA PET/CT imaging (11,13,14,16,23). In a study that analyzed ⁶⁸Ga-PSMA PET/CT positivity in BCR patients after RP, Eiber et al. noted that Gleason score (≥8) was also predictive of higher lesion detection rates (24). However, the predictive value of the time since diagnosis of BCR to scan has not been previously reported. This finding is likely due to progression of disease with longer duration between BCR detection and a PSMA-PET.

Figure 4.

¹⁸F-DCFPyL-PET/CT imaging of a 75-year-old patient with NCCN intermediate-risk prostate cancer (ISUP GG2) treated with a radical prostatectomy. After biochemical recurrence, he underwent a ¹⁸F-DCFPyL-PET/CT at a PSA of 0.5 ng/mL, which showed ¹⁸F-DCFPyL-avid bilateral seminal vesicles without evidence of nodal or distant metastases.

Figure 3.

Nomogram predicting the probability of extra-pelvic disease in BCR patients after radical radiation therapy. Likelihood of the scan positivity was not analyzed due to a small proportion of cohort having negative scan results.

The predictive values of each variable vary when the patient population is stratified by primary treatment. Most studies correlated these variables with positivity of scan, whereas our study analyzed the probability of extra-pelvic disease in addition to scan positivity. Previously our group analyzed a joint cohort of BCR patients from two institutions reporting predictive values for both scan positivity and patterns of disease extension (16). Our current study also stratifies the patient population by primary treatment for analysis, whereas most previous studies limited their patient population to a single primary treatment group (12,13,25) or included the two treatment group patients in one analysis (11). Based on our findings, primary treatment should determine which factors are considered when predicting the chances of a positive scan and the presence of extra-pelvic disease.

Thus, by identifying patients with high probability to result in positive PSMA-PET/CT, we can identify suspicious PCa recurrence(s) and adjust treatment strategies accordingly. Conversely, expensive procedures, often unable to provide definite answers might be avoided when the likelihood of positive PSMA-PET/CT is very low. According to our statistical model, post-surgical patients receiving a score corresponding to a likelihood ≥ 45% (nomogram best cutoff value) may benefit from using the nomogram to determine the need for a PSMA-PET/CT scan. Although the DCA was also constructed for the RT cohort, the net clinical value in getting an ¹⁸F-DCFPyL-PET/CT scan is indeterminate due to the insufficient number of negative scans in our patient cohort. However, the presence of extra-pelvic disease is critical in subsequent management of the patient, and thus, understanding the extent of disease in this patient population could be a deciding factor for obtaining ¹⁸F-DCFPyL-PET/CT scan.

There are some limitations to this study. As previously discussed, the number of patients initially treated with RT with negative test results were insufficient to validate the prediction model. The percentages of negative findings on ¹⁸F-DCFPyL-PET/CT in the RP and RT cohorts were 28.35% (36/127 patients) and 9.72% (7/72 patients), respectively. These negative results may be due to subthreshold size of disease, overlapping urinary bladder within the pelvis, or lack of PSMA expression in dedifferentiated tumors, including, acinar prostate adenocarcinoma and ductal prostate adenocarcinoma (15). No pathology correlation was provided for this study; biopsy sampling of suspected metastases is performed in a limited number of cases and not routinely performed for ethical reasons; namely procedural risks associated with lesions in unsafe locations where biopsy is not feasible (11). However, it is well known that PSMA PET/CT imaging has relatively lower false positive rates compared to other oncologic imaging techniques (26). Finally, our nomogram is developed for ¹⁸F-DCFPyL-PET/CT and may similarly be utilized for other PSMA targeting PET tracers.

Conclusions

In patients with BCR after RP or RT, a nomogram to predict ¹⁸F-DCFPyL PET/CT positivity and presence of disease beyond the pelvis could be used as a guide to select the most suitable candidates for scanning. In both cohorts, PSA and PSAdt were consistently significant predictors for scan positivity and for the presence of extra-pelvic disease. In order to optimally utilize PSMA PET/CT in a health care system, clinicians should be aware of the parameters influencing PSMA-PET positivity in BCR patients and should recognize that differences may exist based on prior treatments.

Figure 6.

¹⁸F-DCFPyL-PET/CT imaging of a patient with NCCN high-risk (ISUP GG4) prostate cancer treated with external beam radiation therapy and long-term androgen deprivation therapy. PSA at the time of scan was 2.9 ng/mL. ¹⁸F-DCFPyL-PET/CT shows focal uptake suspicious of recurrence disease at the left lateral apical-mid prostate, without abnormal distant foci.

List of abbreviations

PCa

Prostate cancer

BCR

biochemical recurrence

PSA

Prostate specific antigen

PSAdt

Prostate specific antigen doubling time

PSMA

prostate specific membrane antigen

PET/CT

positron emission tomography / computed tomography

RP

radical prostatectomy

RT

radiation therapy

AUC

area under the curve

T stage

tumor stage

N stage

nodal stage

NIH

National Institutes of Health

NCI

National Cancer Institute

P

probability

Sdt

standard deviation

DCA

decision curve analysis

TOF

time of flight