Cumulative Radiation Exposures from CT Screening and Surveillance Strategies for von Hippel-Lindau–associated Solid Pancreatic Tumors

Abstract

Purpose

To assess the potential ionizing radiation exposure from CT scans for both screening and surveillance of patients with von Hippel-Lindau (VHL) syndrome.

Materials and Methods

For this retrospective study, abdomen-pelvic (AP) and chest-abdomen-pelvic (CAP) CT scans were performed with either a three-phase (n = 1242) or a dual-energy virtual noncontrast protocol (VNC; n = 149) in 747 patients with VHL syndrome in the National Institutes of Health Clinical Center between 2009 and 2015 (mean age, 47.6 years ± 14.6 [standard deviation]; age range, 12–83 years; 320 women [42.8%]). CT scanning parameters for patients with pancreatic neuroendocrine tumors (PNETs; 124 patients and 381 scans) were compared between a tumor diameter–based surveillance protocol and a VHL genotype and tumor diameter–based algorithm (a tailored algorithm) developed by three VHL clinicians. Organ and lifetime radiation doses were estimated by two radiologists and five radiation scientists. Cumulative radiation doses were compared between the PNET surveillance algorithms by analyses of variance, and a two-tailed P value less than .05 indicated statistical significance.

Results

Median cumulative colon doses for annual CAP and AP CT scans from age 15 to 40 years ranged from 0.34 Gy (5th–95th percentiles, 0.18–0.75; dual-energy VNC CT) to 0.89 Gy (5th–95th percentiles, 0.42–1.0; three-phase CT). For the current PNET surveillance protocol, the cumulative effective radiation dose from age 40 to 65 years was 682 mSv (tumors < 1.2 cm) and 2125 mSv (tumors > 3 cm). The tailored algorithm could halve these doses for patients with initial tumor diameter less than 1.2 cm (P < .001).

Conclusion

CT screening of patients with von Hippel-Lindau syndrome can lead to substantial radiation exposures, even with dual-energy virtual noncontrast CT. A genome and tumor diameter–based algorithm for pancreatic neuroendocrine tumor surveillance may potentially reduce lifetime radiation exposure.

© RSNA, 2018

Online supplemental material is available for this article.

Summary

A tumor size and genotype-based surveillance algorithm could potentially halve the lifetime radiation exposure in patients with von Hippel-Lindau–related solid pancreatic neuroendocrine tumors.

Implications for Patient Care

• ■ CT screening and surveillance of patients with von Hippel-Lindau syndrome can lead to substantial radiation exposures, even with an effective dose of 3–4 mSv for an individual average abdominal-pelvic CT examination, by using a dual-energy virtual noncontrast protocol.

• ■ The use of a tumor size and germline von Hippel-Lindau genotype algorithm may reduce the currently recommended CT scan frequency and the lifetime radiation dose in patients with von Hippel-Lindau syndrome.

Introduction

Von Hippel-Lindau (VHL) syndrome is a familial neoplasia syndrome that occurs in approximately one of 36 000 live births (1). The clinical manifestations of VHL include central nervous system hemangioblastomas, retinal hemangiomas, endolymphatic sac tumors, pheochromocytomas, pancreatic tumors and cysts, including pancreatic neuroendocrine tumors (PNETs), and renal cysts and renal cell carcinomas (1). The condition is caused by a germline pathogenic variation in the VHL tumor suppressor gene and is inherited in an autosomal dominant fashion (2).

Most of the recommended screening guidelines for pancreas tumors involve nonionizing radiation–emitting imaging modalities (sonography and MRI) instead of CT. If repeated during a lifetime, CT can be associated with substantial radiation exposure (Table 1) (3–8). Both CT (9) and MRI (10) have high sensitivity for the depiction of PNETs, and one report (11) suggested superior sensitivity for intravenous contrast agent–enhanced CT over MRI for the depiction of small PNETs. Hence, some organizations recommend the use of CT in asymptomatic adults (3,12). The frequency of imaging examinations varies in the different guidelines, with some recommending annual screening (1,4–7,12,13) and others, biennial evaluations (3,14) depending on whether a pancreatic and/or kidney tumor was already detected in the patient. The medical surveillance of VHL-related PNETs is adapted according to clinical characteristics of the patient and imaging features. MRI is most often used for surveillance of VHL-associated renal cell carcinoma. However, CT is certainly acceptable and may be recommended when needed (3,5,7).

Table 1:

Summary of Recommendations for Screening for Visceral Tumors in Patients with von Hippel-Lindau Syndrome When No Malignant Tumor Is Detected

Note.—NIH = National Institutes of Health, PREDIR = Predispositions aux tumeurs du Rein (predisposition for kidney tumors), VHL = von Hippel-Lindau.

The surveillance protocols for patients with VHL are mainly based on several analyses performed in the National Institutes of Health prospective studies and primarily involve repeated CT scans. Several risk factors for disease progression and for developing metastases have been suggested, including tumor diameter and germline VHL genotype (15). The low risk and high risk of small (15,16) and large tumors (15), respectively, was reported in several publications, whereas to our knowledge other risk factors have not been validated. On the basis of these factors, several algorithms have been suggested for PNET follow-up, but to our knowledge none has been prospectively validated. In a prospective imaging surveillance study conducted at the National Institutes of Health clinical center, patients with pancreatic manifestations of VHL underwent annual CT scanning for PNETs and biennial CT scanning for pancreatic cysts. In our study, we stratified the patients’ risk for requiring surgery (tumor progression and/or developing metastatic PNET) on the basis of their largest tumor diameter and germline VHL genotype (15). Planning a personalized surveillance plan for each patient according to the predicted risk could possibly reduce the number of CT scans.

The purpose of our study was to assess the potential ionizing radiation exposure from CT scans for both screening and surveillance of patients with VHL. The National Institutes of Health VHL cohort data were used to compare alternative screening and surveillance strategies for pancreatic (and renal) tumors. We hypothesized that the use of a low-dose radiation CT protocol, and personalized surveillance algorithms, may lead to substantial reduction in lifetime radiation exposure.

Materials and Methods

Our retrospective study had institutional review board approval and was compliant with the Health Insurance Portability and Accountability Act, and all patients gave written informed consent. The study was supported by the intramural research program of the National Cancer Institute.

Comparison of Radiation Exposures from Alternative Screening Strategies in Asymptomatic Patients

Alternative scenarios of screening strategies for pancreatic tumors were defined on the basis of current guidelines (1,3–8,12,14,17) by clinicians involved in treatment of patients with VHL (A.T., E.K., and W.M.L., with 3, 16, and 36 years of experience, respectively) and radiation scientists, not radiologists (A.B. and N.J. with 17 and 8 years of experience, respectively). The scenarios are based on various combinations of age at start (from 15 to 20 years) and frequency of CT scans (one every year for 2 or 3 years). We also considered scenarios with an increasing frequency of scans in older patients when the radiation risks are believed to be lower and the disease risks are higher (ie, biennial examinations up to age 25 and annual scans afterward).

Imaging Surveillance Protocols among Patients with Pancreatic Tumors

The current National Institutes of Health clinical center protocol for surveillance of pancreatic manifestation or manifestations in patients with VHL who are enrolled in a prospective study evaluating the natural history of VHL-associated pancreatic manifestation or manifestations has been previously described (15). As part of the research protocol, all patients underwent a pancreatic CT protocol. The CT examinations are performed annually in patients with solid pancreatic tumors and every 2 years for patients who have pancreatic cysts regardless of size of cyst or cysts.

As an alternative to this surveillance protocol, two authors (A.T. and E.K.) developed a VHL genotype and tumor diameter–based algorithm on the basis of the 2010–2017 follow-up of 175 cohort members (15). Previous analyses that were based on this cohort are detailed in Table E1 (online), however, none of them compared cumulative radiation exposure between patients’ subgroups. In this previous analysis, it was demonstrated that the surveillance for pancreatic tumors can be optimized on the basis of risk stratification according to tumor size and germline genotype (ie, early surgical intervention for patients with tumor diameter > 3 cm), annual CT scan for those with tumor diameter between 1.2 and 3 cm, with missense, or any pathogenic variation in exon 3 of the VHL gene (high-risk genotype), and biennial or lower surveillance frequency for patients with smaller tumors, or with other VHL genotype (low-risk genotype).

Our current analysis included 124 patients (Fig 1) with a solid pancreatic tumor, a known germline VHL genotype, and who underwent at least two CT scans during follow-up (mean age at inclusion, 50.0 years ± 12.9 [standard deviation]). The patients underwent 381 scans during a median follow-up of 12 months (range, 6–55 months). The interval durations between the scans were 15.0 months ± 6.4 for the entire cohort and 14.8 months ± 4.4, 15.4 months ± 7.1, and 6.9 months ± 5.5 for patients with maximal PNET diameter smaller than 1.2 cm, 1.2–3 cm, and larger than 3 cm, respectively. The intervals between scans among patients with PNET diameter larger than 3 cm were statistically significantly shorter compared with PNET diameter smaller than 1.2 cm and 1.2–3 cm (P = .03 for both comparisons). Among patients with maximal PNET diameter within 1.2–3 cm, interval durations were 14.9 months ± 6.5 and 16.3 months ± 8.3 among patients with high- and low-risk germline VHL genotype, respectively (P = .47).

Figure 1:

Patient selection flowchart. DE = dual energy, VHL = von Hippel-Lindau, VNC = virtual noncontrast protocol.

The 124 patients included in our current analysis were subgrouped according to their pancreatic tumor maximal diameter: smaller than 1.2 cm (61 patients; 49.2%), 1.2–3 cm (60 patients; 48.4%), and larger than 3 cm (three patients; 2.4%). In the proposed algorithm, patients with tumor diameter smaller than 1.2 cm will be followed biennially, and those with tumor diameter larger than 3 cm will be followed annually. Patients with tumor diameter 1.2–3 cm will be further divided by their VHL genotype, to low risk (18 patients; 14.5%) and high risk (42 patients; 33.9%), and will be followed biennially and annually, respectively.

Assessment of Cumulative Radiation Doses

Scanning parameters (manufacturer name and model of the CT scanners, tube current, and volume CT dose index) were retrospectively extracted from Digital Imaging and Communications in Medicine headers, recorded at the National Institutes of Health Clinical Center for nonselected abdominal CT scans performed between 2009 and 2015 in patients with VHL syndrome (W.K. and J.Y., with 3 and 14 years of experience, respectively). In a total of 1391 chest-abdomen-pelvic (CAP) or abdomen-pelvic (AP) scans performed in 747 patients with VHL syndrome (mean age, 47.6 years ± 14.6; age range, 12–83 years; 320 women [42.8%]), 1242 of 1391 scans (89.3%) were performed as three-phase scans, and 149 of 1391 (10.7%) scans were performed with a dual-energy virtual noncontrast (VNC) protocol. Both the three-phase and dual-energy VNC protocols were developed at the National Institutes of Health Clinical Center (L.R.F., with 30 years of experience, and W.K.). Compared with the three-phase CT protocol, the dual-energy VNC protocol halves the radiation exposure by eliminating the precontrast phase and decreasing the tube current with the use of iterative reconstruction (18,19). Eight different models of CT scanners were used. The models included were Brilliance 64 (Philips, Eindhoven, the Netherlands; 904 of 1391 scans [65.0%]), Somatom Definition Flash or AS (Siemens Healthcare, Forcheim, Germany; 355 of 1391 scans [25.5%]), Somatom Force (Siemens Healthcare; 47 of 1391 scans [3.4%]), Somatom Count (Siemens Healthcare; two of 1391 [<0.1%]), Biograph128 (Siemens Healthcare; 58 of 1391 [4.2%]), Aquilion One (Toshiba Medical Systems, Tokyo, Japan; 66 of 1391 [4.7%]), and iCT 256 (Philips; six of 1391 [<0.1%]). Equivalent organ doses and effective dose for each single scan were estimated by radiation scientists (C. Lee, with 14 years of experience, and N.J.) on the basis of volume CT dose index values extracted from the Digital Imaging and Communications in Medicine headers. Effective doses were estimated by using a dedicated computational solution for CT dosimetry, the National Cancer Institute dosimetry system for CT. The calculations were based on pediatric and adult computational human phantoms, coupled with Monte Carlo transport simulation for a reference CT scan model (20,21). Doses associated with standard three-phase CT protocols were applied to both screening and surveillance scenarios, and those associated with the dual-energy VNC protocol were applied to screening scenarios only. For the screening scenarios, doses were accumulated up to age 40 years (ie, the mean age at tumor diagnosis [5,6]) or age 65 years (22). For the surveillance scenarios, doses were accumulated from age 40 years (ie, mean age at detection of first tumor [1,5]) to age 65 years (22).

Parameters of CT Scan Protocols

Three-phase CT protocol.—A nonenhanced scan was performed, followed by injection of intravenous contrast material at a rate of 4 mL/min. The phases were determined by using bolus tracking, with a threshold of 120 HU in the aorta at the level of the diaphragm for the arterial phase, and a venous phase acquisition 70 seconds after completion of the injection.

Dual-energy VNC protocol.—Intravenous contrast material was injected at a rate of 4 mL/min. The phases were determined by using bolus tracking, with a threshold of 120 HU in the aorta at the level of the diaphragm for the arterial phase, and a venous phase acquisition 2 minutes after start of the injection (Fig 2).

Figure 2:

Axial dual-energy virtual noncontrast CT images in a 24-year-old woman with von Hippel-Lindau syndrome. The different appearance of pancreatic neuroendocrine tumor (white arrow), solid renal tumor (arrowhead), and renal cyst (black arrow) are shown between the, A, subtracted virtually noncontrast image and the, B, intravenous contrast–enhanced figure.

Statistical Analysis

Cumulative radiation doses were computed by using statistical software (SAS 9.4, SAS Institute, Cary, NC; and SPSS 20.0, SPSS, Chicago, Ill). Results are expressed as mean ± standard deviation unless otherwise indicated. For group comparisons, the independent Student t test was used to analyze differences in parametric variables, and the χ² test was used to analyze differences in categorical variables. Comparison of cumulative radiation exposure on the basis of diameter-based versus VHL genotype and tumor diameter–based PNET surveillance algorithm were performed by using an analysis of variance that included duration as one of the factors. Comparisons between alternative screening and surveillance scenarios were reported by using descriptive statistics because the exposure scenario parameters were constant and not random variables. The two-tailed P value to indicate statistical significance was set at less than .05.

Results

Radiation Doses from Single CT Scans

CAP scans performed with standard three-phase CT protocols were associated with the highest organ doses (eg, median value, 33 mGy [5th–95th percentiles, 16–39 mGy] per examination to the colon) and the highest effective doses per examination: median value, 25 mSv (5th–95th percentiles, 12–29 mSv) in female patients and 20 mSv (5th–95th percentiles, 10–24 mSv) in male patients (Table 2). Exclusion of the chest from the irradiation field reduced the effective dose by 60%, and the use of dual-energy VNC CT was associated with a further dose reduction of 60%. The lowest median effective doses were 4 mSv (5th–95th percentiles, 2–9 mSv) in female patients and 3 mSv (5th–95th percentiles, 2–7 mSv) in male patients and were associated with an AP examination with the dual-energy VNC protocol.

Table 2:

Effective Dose and Equivalent Organ Doses for Each Single Examination with Standard Three-Phase or Dual-Energy Virtual Noncontrast CT Protocols Used in Patients with VHL

Note.—Data are median values; data in parentheses are 5th–95th percentiles. VHL = von Hippel-Lindau, VNC = virtual noncontrast.

*Calculation of effective doses on the basis of ICRP103 conversion coefficients, which includes hereditary effects and cancer risks at all organs, not only those displayed here as representative of chest-abdominal exposures.

Cumulative Radiation Doses from Alternative CT-based Screening Scenarios

Median cumulative colon dose from annual screening starting at age 15 years ranged from 0.34 Gy (5th–95th percentiles, 0.18–0.75 Gy; dual-energy VNC CT protocol for CAP and AP scans) to 0.89 Gy (5th–95th percentiles, 0.42–1.01 Gy; three-phase CT protocol for CAP and AP scans) by age 40 years and 0.66 Gy (5th–95th percentiles, 0.36–1.48 Gy; dual-energy VNC CT protocol for AP scans) to 1.68 (5th–95th percentiles, 0.82–1.99 Gy; three-phase CT protocol for CAP scans) by age 65 (Fig 3). This highest-exposure screening scenario (one scan every year starting at age 15 years) was associated with median cumulative effective doses ranging from 78 mSv (5th–95th percentiles, 52–182 mSv; dual-energy VNC CT protocol for AP scans) to 520 mSv (5th–95th percentiles, 260–624 mSv; three-phase CT protocol for CAP scans) in male patients by age 40 (Table 3). In female patients, median cumulative effective doses from this screening scenario ranged from 104 mSv (5th–95th percentiles, 52–234 mSv; dual-energy VNC CT protocol for AP scans) to 650 mSv (5th–95th percentiles, 312–754 mSv; three-phase CT protocol for CAP scans, Table 3). The difference in effective doses between male and female patients was mainly attributable to breast cancer risk subsequent to chest irradiation from the CAP scans. Delaying age at screening initiation from 15 years to 18 or 20 years was associated with dose reduction by age 40 years, ranging from 5% and 25%, depending on the frequency of scans and the CT protocol used (Table 3). Decreasing the frequency of scans from once a year to once every 2 or 3 years was respectively associated with 50%–70% dose reduction by age 40 years.

Figure 3:

Cumulative colon dose from CT-based alternative screening strategies for visceral tumors in patients with von Hippel-Lindau. DE-VNC = dual-energy virtual noncontrast.

Table 3:

Cumulative Effective Radiation Dose by Age 40 Years

Note.—Data are median; data in parentheses are 5th–95th percentiles. Calculation of effective doses based on ICRP103 conversion coefficients, which includes hereditary effects and cancer risks at all organs. VNC = virtual noncontrast.

Cumulative Radiation Doses for CT-based Surveillance of Pancreatic Tumors

Among our surveilled study population, seven patients (seven of 124; 5.6%) underwent tumor growth leading to transition from biennial to annual surveillance intervals, after a mean follow-up time of 21.6 months ± 14.5. Three patients (three of 124; 2.4%) underwent resection of the largest pancreatic tumor, which led to a decrease in surveillance frequency from annual to biennial CT scanning after a mean follow-up time from inclusion of 38.2 months ± 19.4.

On the basis of the surveillance algorithm for VHL-associated PNETs and the derived CT scan frequencies and assuming a follow-up between age 40 to 65 years, the lifetime cumulative radiation dose for patients who have PNETs is estimated at 682 mSv for patients with tumors smaller than 1.2 cm in diameter, 667 mSv and 760 mSv for patients with tumor diameter between 1.2 and 3 cm, with versus without missense or exon 3 VHL pathogenic variant, respectively, and 2125 mSv for patients with tumors larger than 3 cm (Fig 4). Comparison of the current surveillance protocol versus the proposed surveillance algorithm showed a statistically significant reduction in lifetime cumulative effective doses of patients with small PNETs, and of those with a largest tumor diameter between 1.2 and 3 cm, and a low-risk VHL genotype (P < .001 for both; Fig 4). Comparison of the difference in cumulative effective doses between patients followed according to the current and proposed surveillance algorithm and tested at different points are detailed in Table 4.

Figure 4a:

Lifetime radiation exposure from CT scans, comparison of two surveillance strategies for patients with von Hippel-Lindau (VHL) and solid pancreatic tumors: current tumor-size based surveillance protocol, as used in the National Institutes of Health, and a proposed tumor's size and genome-based surveillance algorithm. (a) Radiation exposure over time and (b) predicted lifetime radiation exposure among the different risk groups. Two-sided Student t tests, α = .05. The standard deviation from the current surveillance protocol was used for both groups. For lifetime exposure calculations, radiation doses were summed up from age 40 to 65 years. Mis/3 = missense and/or exon 3 germline VHL mutation, non-mis/3 = nonmissense and nonexon 3 germline VHL mutation.

Table 4:

Comparison of Cumulative Effective Radiation Doses in Patients with von Hippel-Lindau and Pancreatic Neuroendocrine Tumors

Note.—Current = current surveillance algorithm, proposed = proposed surveillance algorithm.

*P value was calculated from analysis of variance for each patient subgroup by comparing proposed versus current surveillance algorithm for each group, adjusted by using the Tukey test (all P values for duration were <.001).

^†P value for interaction, <.01.

^‡P value for interaction, <.001.

Figure 4b:

Lifetime radiation exposure from CT scans, comparison of two surveillance strategies for patients with von Hippel-Lindau (VHL) and solid pancreatic tumors: current tumor-size based surveillance protocol, as used in the National Institutes of Health, and a proposed tumor's size and genome-based surveillance algorithm. (a) Radiation exposure over time and (b) predicted lifetime radiation exposure among the different risk groups. Two-sided Student t tests, α = .05. The standard deviation from the current surveillance protocol was used for both groups. For lifetime exposure calculations, radiation doses were summed up from age 40 to 65 years. Mis/3 = missense and/or exon 3 germline VHL mutation, non-mis/3 = nonmissense and nonexon 3 germline VHL mutation.

Discussion

In our study, we show the potentially relatively high radiation exposure for patients who undergo repeated CT scans for screening and surveillance of VHL. High radiation exposure was found even when the dual-energy VNC protocol was used, and it is considered a low-radiation scan protocol. Each single so-called low-dose scan is associated with an average effective dose of 3–4 mSv without imaging the chest or 8–10 mSv with the chest. However, averaged cumulative effective doses by age 40 years could range between 20 and 10 mSv without imaging the chest and between 50 and 260 mSv with the chest, depending on the age at start and the frequency of the scans. With a standard three-phase CT protocol, cumulative doses to age 40 years would exceed 100 mSv without imaging the chest and 200 mSv with the chest in most screening scenarios.

We also compared the cumulative radiation exposure between the current surveillance plan for VHL-related PNETs with a surveillance plan on the basis of germline VHL genotype and PNET diameter. We found that the cumulative radiation exposure may be reduced by more than 40% in patients with small tumors (<1.2 cm diameter) and/or with those with tumor diameter of 1.2–3 cm with the low-risk VHL genotype (no missense or exon 3 VHL pathogenic variant).

The main advantage of the use of MRI or US for the screening and surveillance of PNETs in patients with VHL syndrome is the absence of ionizing radiation exposure. Although MRI was considered inferior to CT to help detect small tumors (11), recent and wider application of diffusion-weighted imaging showed a high diagnostic accuracy compared with older MRI techniques for the depiction of small tumors (<2 cm) (23); therefore, they may be used in both screening and surveillance protocols, in addition to morphologic MRI sequences. The main limitations of the use of MRI is the longer examination time (MRI vs CT, 30–45 minutes vs 5–10 minutes, respectively) and possible contraindications (eg, impaired renal function or implanted electronic devices). However, among young patients with VHL, a low rate of such contraindications is expected and the use of diffusion-weighted imaging allows for a reduction in the use of the contrast material. For these reasons, greater efforts must be directed toward the use of MR protocols even without the use of contrast material. This is further supported by the fact that MRI is the preferred method for screening and surveillance of renal cysts and carcinoma because of its accuracy and lower risk for kidney injury, and because there is no ionizing radiation exposure. Moreover, even when intravenous contrast-enhanced CT is used, it should be noted that the late arterial phase is sensitive and usually sufficient for the depiction of PNETs, possibly obviating a three-phase examination (24).

In terms of the imaging techniques to be used, the current guidelines for screening in asymptomatic patients with VHL differ among various organizations and countries. Organizations from the United States usually recommend CT-based screening in adults, whereas European organizations typically recommend the use of MRI for screening purposes unless a solid tumor has been detected and/or for preoperative assessment. The VHL Alliance, a nonprofit organization that serves as the main resource of VHL-related information, revised its screening recommendation in 2016 from CT to MRI, stating that “CT should be avoided for all pre-symptomatic people, and should be reserved for occasions when it is truly needed to answer a diagnostic question” (10). International expert consensus for the use of imaging methods with no or lower ionizing radiation exposure may thus be warranted to improve patient care.

Patients with VHL-associated pancreatic tumors require lifelong surveillance, including yearly imaging or imaging every 2–3 years, depending on the type of pancreatic abnormal finding (mass and/or cyst, respectively) in addition to surveillance tests required for the other disease manifestations (25,26). The natural history of pancreatic tumors is usually of the indolent type, with a low event rate and slow tumor growth, if at all (15,27). Hence, a personalized approach is required for identifying high-risk pancreatic tumors that require closer follow-up and for reducing the radiation exposure associated with follow-up for those with a lower risk for disease progression. On the basis of our data, we can now offer a large subset of our patients less frequent follow-up and CT scans compared with the previous, recent surveillance policy, enabling a substantial reduction in lifetime cumulative radiation exposure.

To our knowledge, no previous study has assessed the radiation exposures for screening and surveillance of VHL manifestations. However, our current analysis had limitations. First, our single-institution study may not reflect practices in other facilities. Second, although several hereditary cancer syndromes have been associated with a higher sensitivity to radiation-induced malignancies (28,29), to our knowledge, no previous study has specifically investigated the effect of radiation exposures in patients with VHL. The cumulative radiation doses in this population are potentially high, and a number of studies have directly demonstrated excess cancer risks from repeated diagnostic radiation exposures in this dose range (30,31). With current evidence, we can assume that the risks of developing radiation-related malignancies (and potentially noncancer events such as circulatory diseases) are similar to those of the general population (28,32); however, higher risks or earlier occurrence of diseases in patients with VHL are also plausible. Consequently, considering the potential for high cumulative radiation doses from a lifetime of screening and surveillance for pancreatic tumors, CT exposures should be reduced as much as possible. Nonionizing imaging modalities should be used, or CT scans used less frequently. The use of low-radiation-dose CT protocols and avoidance of exposing large body areas (in particular, excluding the gonads and chest from the imaging volume) are efficient options for dose reduction (33). Third, in the previous surveillance policy for VHL-related pancreatic manifestations used at the National Institutes of Health Clinical Center, the frequency of CT scans varied depending on the type of mass (ie, solid and/or cystic). However, in the VHL genotype and tumor diameter-based algorithm, pancreatic cysts were not included, thus limiting the comparison between them. Forth, this was a retrospective analysis that has inherent limitations that should be considered when interpreting the analysis results. Fifth, small renal masses (<1.5 cm) may be difficult to characterize because of an oscillation of the attenuation values of 5–20 HU between the virtual and real nonenhanced images (34). Moreover, in the characterization of solid pancreatic masses at dual-energy VNC, artifacts may lead to improper subtraction of existing masses (35). Finally, dual-energy VNC often cannot be used in patients with large body habitus or with large abdominal circumference because of the low penetration power of the 80-kVp beam and the limited field-of-view range of that particular tube.

On the basis of our current analysis, we recommend the use of a dual-energy VNC CT protocol in patients with VHL and to personalize the surveillance of patients with VHL and pancreatic manifestations based on their tumor diameter and germline VHL genotype to allow reduced cumulative radiation dose in low-risk patients. Moreover, although not directly derived from the current analysis, we suggest re-evaluating the role of MRI in the screening and surveillance of patients with pancreatic manifestations of VHL, and to consider the use of MRI as a substitute to CT when available.

In conclusion, CT screening of patients with VHL syndrome for visceral tumors can lead to substantial radiation exposures. Even with effective radiation dose of 3–4 mSv for AP examination at dual-energy VNC CT, relatively high lifetime doses to radiation-sensitive organs can result from annual screening. When it is feasible, alternative imaging procedures, particularly MRI, should be considered. With pancreatic manifestations of VHL, a tailored approach for each patient on the basis of the germline VHL genotype and the size of the pancreatic tumors may further reduce the lifetime radiation exposure.