Skip to main content

Skyshine and Roof Shielding for Imaging Facilities

By Troy Zhou, PhD, DABR, DABSNM
• 19 min read

Skyshine is radiation that leaves a facility through a thinly shielded roof, scatters in the air overhead, and comes back down to ground level or to an occupied floor of a taller building nearby — a path a side-wall calculation never sees. Roof and ceiling barriers matter for PET/CT and nuclear medicine suites, where 511 keV photons from injected patients reach the floor directly above, and for megavoltage linac vaults, where the classic NCRP skyshine estimate is a conservative screening tool that modern measurement and Monte Carlo work show tends to overpredict. 161113

Shielding design usually begins with walls, and for good reason — barriers facing a beam or a nearby source carry most of the dose. But radiation does not travel only in straight lines through walls. Photons that pass up through a comparatively light roof scatter off the atmosphere and can return to occupied areas, and photons from a source inside the building reach the floor above through the ceiling slab. 16 This article explains the skyshine mechanism, how roof and floor-above barriers are evaluated for 511 keV and megavoltage facilities, a worked floor-above calculation, why the legacy skyshine formula should be used cautiously, practical design tips, and the regulatory context. DRPS provides this analysis through its radiation shielding design service.

Introduction

The overhead problem has three distinct dose pathways, and a complete shielding review accounts for all three. Direct transmission sends primary, scatter, or leakage radiation straight up through the roof; skyshine sends radiation up through the roof and back down via air scatter; and groundshine scatters radiation across from the ground and nearby surfaces. 13

Pathway What it is Where it dominates Governing reference
Direct roof/ceiling transmission Primary, leakage, and scatter passing up through the barrier Floor directly above a PET/CT or nuclear medicine room (511 keV patient source); directly above a vault AAPM TG-108 6; NCRP 147/151 12
Skyshine Radiation up through a light roof, scattered by air, returning to grade or a nearby taller building Megavoltage linac vaults with lightly shielded roofs NCRP 151 1; NCRP 144 3
Groundshine Radiation scattered from the ground and surrounding surfaces Adjacent occupied points near a vault or suite NCRP 151 1

The two most common medical settings are different in character. In a PET/CT or nuclear medicine suite, the 511 keV annihilation photons from the injected patient are the dominant overhead source, and the key question is the barrier for the occupied floor directly above — a direct-transmission problem solved with TG-108 methods. 6 In a megavoltage linac vault, the roof is often lightly shielded because no one stands on it, so skyshine becomes the pathway of interest, especially toward a nearby taller building. 111 For the wall-shielding companions to this overhead view, see our PET/CT shielding calculations guide and linac vault shielding with NCRP 151.

Topic Explanation

What is skyshine?

Skyshine is the component of dose delivered to a point outside or above a facility by radiation that first travels upward through a weak roof barrier and is then scattered back toward the ground by the air. Because air is a diffuse scatterer, the returning radiation arrives over a broad area, and the dose peaks not directly beside the building but some distance away. 111

Monte Carlo analysis of medical accelerators gives a useful picture of where and what skyshine is. The maximum skyshine dose occurs at a horizontal distance of roughly twice the roof height above the isocenter, not at the building wall, and the mean energy of the returning scattered photons is only about 0.15 MeV — far softer than the primary beam — nearly independent of the machine energy and geometry. 1112 At intermediate distances the skyshine dose falls roughly as one over the distance, more slowly than the inverse-square falloff of a direct source, which is why a nearby taller building can matter even when it is not immediately adjacent. 11

What radiation sources matter overhead?

For a PET/CT or nuclear medicine suite, the overhead source is the injected patient and handled activity, emitting 511 keV photons. The floor above an uptake room, scanner room, or hot lab can receive meaningful dose, so the ceiling slab is treated as a secondary barrier under TG-108 methodology just like a wall. 6 For a megavoltage vault, the sources are the primary beam directed upward, plus leakage and patient scatter, passing through the roof and scattering in air. 13

For the broad-beam transmission physics these barriers depend on, see buildup factor and broad-beam gamma shielding and the general lead shielding design principles.

Design goals and occupancy

Roof and floor-above barriers are designed to the same weekly goals as walls: a common objective is 0.1 mGy per week (about 5 mGy per year) for controlled areas and 0.02 mGy per week (about 1 mGy per year) for uncontrolled areas, consistent with the public and occupational limits of 10 CFR Part 20. 146 Occupancy matters especially overhead: an occupied office on the floor above carries an occupancy factor near 1, while an unoccupied roof or a mechanical penthouse carries a much smaller factor, and the occupancy of an adjacent taller building is what makes skyshine a design driver or a non-issue. 1

Key Technical Principles

The 511 keV floor-above barrier (TG-108)

For a PET source, the overhead barrier is a direct-transmission calculation. The unshielded dose rate at an occupied point above the source, at distance , is:

where is the 511 keV effective dose-rate constant and the activity present. AAPM TG-108 provides an effective dose-rate constant for F-18 in the patient of about after accounting for patient self-attenuation, and recommends a larger value of about (no patient attenuation) as a conservative choice for shielding. 6 The required barrier transmission for a weekly design goal , occupancy , and total unshielded weekly dose is:

and the thickness follows from the tenth-value layer for 511 keV:

A commonly used lead TVL at 511 keV for illustrative design is about 1.6 cm. 6

Worked floor-above example

Consider an uptake room with an occupied office on the floor directly above. Assumptions:

  • Radiopharmaceutical: F-18 FDG; administered activity 370 MBq per patient.
  • Patients per week: 30; time near the barrier: 1 hour per patient.
  • Effective dose-rate constant: . 6
  • Vertical distance from patient to the occupied point one floor up: 3.5 m.
  • Occupancy of the office above: .
  • Weekly design goal (uncontrolled): . 16

The unshielded dose rate is:

The weekly unshielded dose is:

The required transmission is:

and the lead-equivalent thickness is:

About a centimeter of lead equivalent would meet the goal for this occupied floor above — but much of that may already be supplied by the structural concrete slab, which is why the ceiling must be evaluated in three dimensions and not assumed adequate. A complete design would also add contributions from other 511 keV source locations and verify the result by survey. 6

The skyshine estimate for megavoltage vaults

For a megavoltage vault, NCRP Report No. 151 provides an analytical skyshine estimate. In its general form the skyshine dose-equivalent rate is proportional to the beam output and the solid angle subtended by the beam, and inversely proportional to the product of the squared vertical and horizontal distances:

where is the beam output with the beam directed upward, is the solid angle (steradians) subtended by the collimator opening, is the vertical distance from the target to a point just above the roof, and is the horizontal distance from the isocenter to the point of interest. 13 The leading coefficient and its units must be taken from the current NCRP Report No. 151 edition rather than quoted loosely, because published renderings differ by orders of magnitude depending on the dose and output units chosen. The solid angle itself depends on field shape: for a square beam the pyramidal solid angle is about 1.27 times the circular (conical) value, so the correct geometry must be used. 10

This equation should be treated as a conservative screening estimate, not a validated prediction. Direct measurements on 6 MV and 10 MV accelerators found that the NCRP skyshine model did not reproduce the measured ground-level and roof dose rates, 79 and a separate measurement study across eight accelerators reported differences between measured and empirical skyshine of one or more orders of magnitude. 8 Monte Carlo analyses likewise show the standard formula is inaccurate — skyshine scales more nearly with beam field area than with — and provide replacement scaling relationships. 1112 A 2025 review of megavoltage skyshine studies concluded that NCRP 151 tends to overestimate skyshine. 13 The practical consequence is that the NCRP estimate is useful as a first, conservative screen; a tight or contested design should be confirmed with measurement or modern modeling, and a useful rule of thumb from the Monte Carlo work is that if the roof is shielded enough to avoid a "high radiation area" designation, photon skyshine is usually negligible. 11

Roof thickness and the energy surprise

Roof barriers at megavoltage energies are substantial where occupancy demands it: NCRP 151 tabulates primary-barrier tenth-value layers on the order of tens of centimeters of concrete — approximately 37 cm of concrete at 6 MV, rising with energy. 1 But for a thin roof the energy dependence of skyshine inverts intuition: Monte Carlo work shows a 6 MV beam can produce more skyshine than an 18 MV beam because the lower-energy photons scatter through larger angles back toward the ground, with the crossover near about 0.5 m of concrete. 1112 A designer cannot assume that a higher-energy machine is always the worse skyshine case.

Clinical Impact

Overhead shielding decisions shape building layout, floor stacking, and cost long before the first patient — and getting them wrong is expensive to fix. A PET/CT suite placed under occupied offices, or a linac vault with occupied space or a taller building nearby, converts an overhead pathway from an afterthought into a design driver. 16

For PET/CT and nuclear medicine, the floor above is the common oversight. The injected patient is a mobile 511 keV source that spends uptake and imaging time in rooms whose ceilings may have been treated as structural slabs rather than radiation barriers. Because 511 keV photons are penetrating and the patient activity is in the hundreds of megabecquerels, a full-occupancy office directly above an uptake room can require added shielding, as the worked example shows. 6 Vertical stacking of low-occupancy spaces above and below high-activity rooms is the most economical fix, and it is only available before the layout is fixed. 6

For megavoltage vaults, skyshine toward a taller neighboring building is the classic surprise, and the measurement and modeling literature carries a consistent message: the legacy estimate is conservative, so a vault that "fails" a screening skyshine calculation may still be acceptable on measurement, while a vault that passes the screen is very likely safe. 7813 The right posture is to screen conservatively, design with margin, and verify by survey rather than trusting any single formula.

Practical Optimization Tips

Evaluate the facility in three dimensions

Plan-view adjacency is not enough. Identify every occupied space above and below each radiation room, and every taller occupied building within the skyshine range, and assign realistic occupancy to each. The floor above a PET uptake room and a nearby office tower are the points a wall-only review misses. 16

Use the right method for the source

Use TG-108 direct-transmission methods for the 511 keV floor-above barrier in PET/CT and nuclear medicine, and the NCRP 151 skyshine estimate for the megavoltage vault roof, understanding the latter as conservative. 6111 Applying a wall formula to a skyshine path, or a skyshine formula to a direct ceiling path, gives the wrong answer.

Treat the NCRP skyshine number as a screen, then verify

Because the legacy skyshine equation can overpredict, use it to decide whether skyshine is plausibly a concern, then confirm a tight design with measurement or Monte Carlo rather than building to the conservative estimate alone. 111213 Document which method produced which number.

Credit — but verify — the structural slab

A concrete floor slab already provides attenuation, so the added lead for a floor-above barrier is often small, as in the worked example. But credit the slab only for its verified density and thickness, and account for penetrations, conduit, and construction joints, which are the real weak points. 16

Survey the as-built condition

No overhead shielding design is complete until a post-construction survey confirms measured dose rates against the design goals in the occupied floor above and at the skyshine-relevant exterior points, with an energy-appropriate, calibrated instrument. 16 For the measurement step, see our guides to the radiation protection shielding survey and choosing the right radiation survey meter.

Common pitfalls to avoid

  • Treating the ceiling as structure, not a barrier. The floor above a 511 keV source can need shielding. 6
  • Building to the raw NCRP skyshine number. It tends to overestimate; screen, then verify. 1113
  • Assuming higher energy is always worse for skyshine. For thin roofs, lower energy can give more. 1112
  • Using the wrong solid angle. Square fields need the pyramidal solid angle, about 1.27 times the conical value. 10
  • Skipping the post-construction survey. The calculation is a prediction; the survey is the proof. 1

Regulatory Considerations

Roof, ceiling, and skyshine shielding must satisfy the same dose limits and design goals as any other barrier, documented so the analysis is defensible during plan review and inspection. The methodology differs by facility type, but the design goals and the authority-having-jurisdiction requirements do not.

Key frameworks to reference:

  • NCRP Report No. 151 — structural shielding design for megavoltage radiotherapy facilities, including the skyshine methodology and primary-barrier transmission data. 1
  • NCRP Report No. 147 — structural shielding design for medical X-ray imaging facilities, the diagnostic-imaging counterpart for roof and floor barriers. 2
  • NCRP Report No. 144 — radiation protection for particle accelerator facilities, which treats photon and neutron skyshine and overhead shielding. 3
  • AAPM Task Group 108 — PET and PET/CT shielding requirements, the basis for the 511 keV floor-above calculation. 6
  • IAEA Safety Reports Series No. 47 — radiation protection in the design of radiotherapy facilities, with roof and skyshine worked examples. 5
  • 10 CFR Part 20 — the NRC dose limits that set the design goals for occupational and public exposure. 4

X-ray-producing machines, including linacs and CT, are regulated by state radiation-control programs, while byproduct material such as the PET radionuclides falls under NRC or Agreement State rules; a facility must confirm which authority applies and verify requirements before relying on any shielding assumption. For the compliance picture, see common radiation safety violations and how to avoid them and workload, use, and occupancy factors in shielding.

Frequently Asked Questions (FAQs)

What is skyshine in radiation shielding?

Skyshine is radiation that escapes through a thinly shielded roof, scatters off air molecules above the building, and returns to ground level some distance away or to an occupied upper floor of an adjacent taller structure. Because the path goes up and back down rather than straight through a wall, skyshine can reach areas a side-wall calculation would miss. 111

Which facilities need to consider roof and skyshine shielding?

Megavoltage linac vaults are the classic skyshine concern. PET/CT and nuclear medicine suites need roof and floor-above evaluation because the injected patient is a 511 keV source that can expose the floor directly above. Any facility with occupied space above a radiation room, or a taller building nearby, should have the overhead path evaluated. 16

How is 511 keV roof or floor-above shielding calculated?

The injected patient is treated as a 511 keV source using AAPM Task Group 108 methods: estimate the activity present, compute the unshielded dose at the occupied point one floor up with the dose-rate constant and the inverse-square law, apply occupancy, solve for the required transmission, and convert to lead or concrete using the 511 keV tenth-value layer. 6

Is the NCRP 151 skyshine equation accurate?

It is best treated as a conservative screening estimate. Measurements and Monte Carlo studies of megavoltage accelerators have found the classic formula can disagree with measured dose rates, sometimes by close to an order of magnitude, and a recent review concluded it tends to overestimate skyshine. Use it as a first screen and confirm a tight design by measurement or modeling. 781113

Does low photon energy produce more skyshine than high energy?

For a thin roof it can. Monte Carlo work shows a 6 MV beam can produce more skyshine than an 18 MV beam, because lower-energy photons scatter through larger angles back toward the ground, with the crossover near about half a meter of concrete roof thickness. 1112

What is groundshine and how is it different from skyshine?

Groundshine is radiation scattered from the ground and surrounding surfaces, contributing to dose at nearby occupied points, whereas skyshine comes down from the air over the roof. Both are secondary, scattered pathways evaluated alongside direct barrier transmission. 13

Key Takeaways

  • Overhead has three pathways. Direct roof/ceiling transmission, skyshine, and groundshine are each evaluated in a complete review. 136
  • PET/CT needs the floor above. The injected 511 keV patient can expose the occupied floor directly overhead; use TG-108 methods. 6
  • Skyshine drives megavoltage vaults. Light vault roofs let high-energy photons scatter in air and return to grade or a taller neighbor. 111
  • The NCRP skyshine formula overpredicts. Measurement and Monte Carlo show it is conservative; screen with it, then verify. 7813
  • Energy can invert for thin roofs. A 6 MV beam can give more skyshine than 18 MV below about 0.5 m of concrete. 1112
  • Verify by survey. The calculation predicts; the post-construction survey proves. 16

Conclusion

Radiation does not respect the plan view. For PET/CT and nuclear medicine, the 511 keV patient can expose the floor directly above, a direct-transmission problem solved cleanly with TG-108 methods and often met by the structural slab plus a modest added barrier. For megavoltage vaults, skyshine sends softened photons up through a light roof and back down to grade or a taller neighbor, and the legacy NCRP estimate — while a useful conservative screen — overpredicts according to the measurement and Monte Carlo record, so it should be screened with and then verified. In both cases the discipline is the same: evaluate the facility in three dimensions, use the right method for the source, design with margin, and confirm the result with a survey. 161113

How DRPS Can Help

Diagnostic Radiation Physics Services helps imaging and therapy facilities translate overhead shielding requirements into practical, documented designs. This can include radiation shielding design for roof, ceiling, and floor-above barriers, 511 keV floor-above evaluation for PET/CT and nuclear medicine, skyshine screening for megavoltage vaults, post-construction verification surveys, and RSO program guidance and medical physics consulting performed by board-certified medical physicists.

DRPS supports facilities across our service locations, including Florida, Maryland, Virginia, Washington DC, California, Nevada, New York, Pennsylvania, New Jersey, and Delaware.

A strong overhead shielding review is not about adding lead to every ceiling. It is about finding the one occupied space above or beside the room that the wall calculation forgot.

Related Resources

References

  1. National Council on Radiation Protection and Measurements. Structural Shielding Design and Evaluation for Megavoltage X- and Gamma-Ray Radiotherapy Facilities. NCRP Report No. 151. Bethesda, MD: NCRP; 2005. ncrponline.org
  2. National Council on Radiation Protection and Measurements. Structural Shielding Design for Medical X-Ray Imaging Facilities. NCRP Report No. 147. Bethesda, MD: NCRP; 2004. ncrponline.org
  3. National Council on Radiation Protection and Measurements. Radiation Protection for Particle Accelerator Facilities. NCRP Report No. 144. Bethesda, MD: NCRP; 2003. ncrponline.org
  4. U.S. Nuclear Regulatory Commission. 10 CFR Part 20: Standards for Protection Against Radiation. ecfr.gov
  5. International Atomic Energy Agency. Radiation Protection in the Design of Radiotherapy Facilities. Safety Reports Series No. 47. Vienna: IAEA; 2006. iaea.org
  6. Madsen MT, Anderson JA, Halama JR, et al. AAPM Task Group 108: PET and PET/CT shielding requirements. Med Phys. 2006;33(1):4-15. doi:10.1118/1.2135911. doi.org
  7. Elder DH, Harmon JF, Borak TB. Skyshine radiation resulting from 6 MV and 10 MV photon beams from a medical accelerator. Health Phys. 2010;99(1):17-25. doi:10.1097/HP.0b013e3181d3d828. doi.org
  8. de Paiva E, da Rosa LAR. Skyshine photon doses from 6 and 10 MV medical linear accelerators. J Appl Clin Med Phys. 2012;13(1):3671. doi:10.1120/jacmp.v13i1.3671. doi.org
  9. Gossman MS, McGinley PH, Rising MB, Pahikkala AJ. Radiation skyshine from a 6 MeV medical accelerator. J Appl Clin Med Phys. 2010;11(3):3032. doi:10.1120/jacmp.v11i3.3032. doi.org
  10. Gossman MS, Pahikkala AJ, Rising MB, McGinley PH. Providing solid angle formalism for skyshine calculations. J Appl Clin Med Phys. 2010;11(4):3286. doi:10.1120/jacmp.v11i4.3286. doi.org
  11. McDermott PN. Photon skyshine from medical linear accelerators. J Appl Clin Med Phys. 2020;21(3):108-114. doi:10.1002/acm2.12833. doi.org
  12. McDermott PN. Medical linac photon skyshine: Monte Carlo calculations and a methodology for estimates. J Appl Clin Med Phys. 2022;23(4):e13543. doi:10.1002/acm2.13543. doi.org
  13. Molazadeh M, Ghiasi H. A review of photon beam skyshine studies from medical megavoltage accelerators. Appl Radiat Isot. 2025;219:111716. doi:10.1016/j.apradiso.2025.111716. doi.org