Digital Radiography Detectors: Direct vs Indirect

Flat-panel digital radiography detectors fall into two families defined by a single physics choice: direct-conversion photoconductors that turn absorbed X-rays straight into charge, and indirect-conversion scintillators that first make light and then convert that light to charge. That choice ripples through spatial resolution, image noise, detective quantum efficiency, and dose efficiency, so understanding it is the foundation for specifying, acceptance-testing, and troubleshooting any DR system.¹²⁶

Direct-conversion detectors, most commonly built from amorphous selenium (a-Se), collect charge with almost no lateral spread and therefore deliver very sharp images. Indirect-conversion detectors, most commonly cesium iodide (CsI) or gadolinium oxysulfide coupled to an amorphous-silicon (a-Si) photodiode array, add an optical step that can blur fine detail slightly but often absorbs the beam more efficiently. Neither family is universally "better"; the right detector depends on the imaging task, the beam quality, and how the system is calibrated and maintained.⁵⁶

Introduction

Digital radiography replaced screen-film with electronic detectors that read out a latent image directly, but two very different physical mechanisms are used to do it. A medical physicist who understands those mechanisms can explain why one chest unit looks razor-sharp while another produces lower-noise images at the same dose, and can hold acceptance testing and quality control to the right standard for each design.³⁶

The distinction matters clinically and economically. Detector conversion type influences limiting spatial resolution, low-contrast detectability, dose efficiency, susceptibility to specific artifacts, and even the practical robustness of a portable panel dropped on a hospital floor. Purchasing decisions, protocol optimization, and dose management all lean on these fundamentals.⁶⁷

This article walks through what "direct" and "indirect" conversion actually mean, the physics that links conversion mechanism to measurable image-quality metrics, a worked detective-quantum-efficiency example, the clinical consequences, practical QC guidance, and the regulatory context a facility should keep in view. Throughout, the goal is to connect detector physics to decisions a radiology department makes every day.

Topic Explanation

A digital radiography detector converts the spatial pattern of X-rays transmitted through the patient into a digital image, and the conversion can happen in one step or two. In both designs, a thin-film-transistor (TFT) active-matrix array reads out the collected charge pixel by pixel, but what happens above that array differs fundamentally.³⁴

Direct conversion

In a direct-conversion detector, a thick layer of X-ray photoconductor — almost always amorphous selenium — sits on top of the TFT array under a high applied electric field. When an X-ray is absorbed in the a-Se, it liberates electron-hole pairs. The applied field sweeps that charge along nearly straight field lines to the collection electrode of the pixel directly beneath the interaction, with very little sideways migration. Because the charge is not allowed to wander, the spatial information is preserved almost perfectly, and the detector behaves close to an ideal sampling aperture set by the pixel size.¹⁸

Indirect conversion

In an indirect-conversion detector, an X-ray is first absorbed in a scintillator, which emits a burst of visible light. That light is then detected by an a-Si photodiode at each pixel and converted to charge. Two scintillator geometries are common: unstructured (turbid) phosphors such as gadolinium oxysulfide (Gd₂O₂S), in which light spreads laterally as it diffuses toward the photodiodes, and structured cesium iodide grown in fine needle-like columns that act as light guides, channeling photons downward and limiting lateral spread. Structured CsI recovers much of the sharpness that would otherwise be lost to light diffusion, which is why it dominates high-performance indirect DR.²⁶

Computed radiography for context

Cassette-based computed radiography (CR) uses a photostimulable storage phosphor that traps energy from absorbed X-rays and releases it as light only when scanned by a laser in a separate reader. CR is an indirect, light-mediated technology, but its offline readout and light spread within the phosphor generally place its detective quantum efficiency and resolution below integrated flat-panel DR. Understanding CR's limits is part of understanding why hospitals migrated to flat-panel DR.²⁵⁶

For the metric that quantifies how efficiently each of these detectors uses the X-ray beam, see our explainer on detective quantum efficiency in digital radiography.

Key Technical Principles

The conversion mechanism controls three linked image-quality metrics — the modulation transfer function (MTF), the noise power spectrum (NPS), and the detective quantum efficiency (DQE) — and DQE is the one that ties image quality to patient dose.⁵⁶

Signal, sharpness, and the MTF

The MTF describes how faithfully a detector reproduces spatial detail as a function of spatial frequency. Direct-conversion a-Se, with negligible lateral charge spread, typically shows an MTF close to the theoretical curve of the pixel aperture, so its high-frequency response is excellent. Indirect detectors lose some MTF to light spreading; structured CsI limits that loss, while unstructured phosphors show a lower MTF. In a controlled comparison of commercial systems, the direct-conversion detector's presampled MTF was significantly higher than that of the indirect systems and very close to the ideal pixel-aperture function.⁵

Noise and the Swank factor

Detected quantum noise is shaped by how uniformly each absorbed X-ray produces signal. In a scintillator, the number of light quanta produced and detected per X-ray fluctuates, adding an extra noise term. The Swank factor $A_S$ quantifies this loss using the moments $M_i$ of the distribution of detected quanta per interacting X-ray:

A perfectly uniform response gives $A_S=1$; real scintillators fall below that, degrading DQE. Direct-conversion a-Se avoids the optical-gain stage and is not subject to the same light-statistics penalty, one reason its low-frequency noise behaves so cleanly.¹⁶

Dose efficiency and the DQE

DQE is the fraction of the incident X-ray signal-to-noise ratio (squared) that survives detection — the single best measure of dose efficiency. In the frequency domain, the standardized form used for detector characterization is:

where $\mathrm{NNPS}(f)$ is the normalized noise power spectrum and $q$ is the incident photon fluence (photons per unit area) for the specified beam. At zero frequency the DQE simplifies to a product of physical efficiencies:

where $\eta$ is the quantum detection efficiency (the fraction of incident X-rays absorbed by the converter) and $A_S$ is the Swank factor. This shows why a thick, strongly absorbing converter matters: a CsI layer engineered to absorb a large fraction of the beam raises $\eta$, and if it maintains a high Swank factor it delivers high DQE(0).⁵⁶

Worked DQE example

Consider an indirect CsI/a-Si detector with a quantum detection efficiency of $\eta =0.70$ at a typical chest beam quality and a Swank factor of $A_S=0.90$. Its zero-frequency DQE is approximately:

Now compare a direct a-Se detector of the same nominal thickness whose absorption at the same beam quality gives $\eta =0.55$, but with an effectively higher Swank behavior $A_S\approx 0.98$ because it avoids optical statistics:

In this illustrative case the indirect detector wins at low frequency on absorption alone, while the direct detector's superior MTF lets its DQE hold up better toward high spatial frequencies. The trade-off between low-frequency dose efficiency and high-frequency sharpness is exactly the design tension between the two families, and it is why the "better" detector is task-dependent.⁵⁶

Comparison table

Values are qualitative and device-dependent; a specific unit must be characterized by measurement under standardized conditions rather than assumed from its category.⁵⁶¹⁰

Clinical Impact

Detector conversion type changes what a radiologist sees and what a patient receives. A direct a-Se system can resolve fine trabecular detail and microcalcification-scale structure exceptionally well, which is one reason a-Se became widespread in digital mammography, where high-frequency performance and low noise at a soft beam quality are decisive.⁶⁸

For general projection radiography — chest, abdomen, skeletal, and portable work — the balance often favors indirect structured CsI. Published comparisons have found that indirect-conversion CsI DR detectors offer strong physical image quality and dose efficiency for standard projection radiography, and clinical evaluations have supported those laboratory findings. That dose efficiency can be spent two ways: better images at the same exposure, or equivalent images at lower exposure, supporting ALARA for the patient.⁶

The detector also interacts with the exposure index, the system's feedback signal for how much radiation actually reached the detector. Because digital systems produce a usable image over a very wide exposure range, over- and under-exposure are not obvious from image brightness the way they were with film. Without a well-managed exposure index, techniques tend to creep upward over time — "dose creep" — silently increasing patient exposure. The detector's efficiency sets the floor, but disciplined exposure-index monitoring keeps a department operating near it.⁹

Artifact patterns differ too. Indirect detectors can show light-spread-related blur and scintillator non-uniformities; direct a-Se can be sensitive to temperature, humidity, and high-exposure lag or ghosting behavior. Knowing the technology tells a physicist what artifacts to look for during troubleshooting. For related detector-uniformity testing, see our guide to flat-panel detector uniformity QC.

Practical Optimization Tips

Match the detector to the task, then hold it to a measured standard. A few principles help facilities and physicists get the most from either technology:

1. Specify by measured DQE and MTF, not marketing category. Ask vendors for MTF, NNPS, and DQE curves measured under IEC 62220-1-1 conditions, and confirm them at acceptance. Two detectors of the same "type" can differ substantially.⁵¹⁰
2. Weigh the clinical task. For high-detail work such as mammography, prioritize the high-frequency MTF where direct a-Se excels. For general and chest radiography at higher energies, prioritize absorption and low-frequency DQE where structured CsI often leads.⁶⁸
3. Calibrate and monitor the exposure index. A correctly configured, regularly checked exposure index is the single most effective defense against dose creep, independent of detector type. Establish target ranges and review outliers.⁹
4. Track environmental conditions for a-Se. Direct-conversion selenium detectors can be sensitive to temperature and humidity; keep them within manufacturer limits and watch for lag or ghosting after high-exposure images.¹⁸
5. Build detector-specific QC. Use uniformity, signal transfer property, limiting resolution, and artifact evaluations tuned to the detector's known failure modes, on the schedule described by AAPM Task Group guidance and the manufacturer.⁹¹¹
6. Re-baseline after service. Detector replacement, gain recalibration, or firmware updates can shift image quality and the exposure-index calibration; re-verify constancy afterward rather than assuming the prior baseline still applies.⁹

Regulatory Considerations

Digital radiography sits under overlapping oversight: federal equipment standards, state radiation-control rules for the X-ray machine, and consensus medical-physics standards for how detector performance is measured.

• X-ray machine performance. The X-ray generator and tube are regulated as radiation-producing machines under state radiation-control programs, with federal performance standards for diagnostic equipment codified at 21 CFR 1020.31. State programs — including Florida's under Chapter 64E-5, F.A.C. — generally require registration, periodic performance evaluation, and records for diagnostic radiographic equipment.¹²
• Detector characterization. IEC 62220-1-1:2015 is the international standard method for determining detector DQE (via the signal transfer property, MTF, and normalized NPS) under defined beam qualities such as RQA5, and is the reference many acceptance tests and published comparisons follow.¹⁰
• Exposure-index management. AAPM Task Group 116 defined a standardized exposure indicator for digital radiography so facilities can monitor detector exposure consistently, and AAPM Task Group 151 provides a framework for ongoing quality control in digital radiography, including exposure-index review, uniformity, and artifact testing.⁹¹¹

Jurisdiction matters for who reviews and enforces these requirements. X-ray machines are regulated by the state radiation-control program (in Florida, the Department of Health, Bureau of Radiation Control); a board-certified medical physicist typically performs the acceptance and annual performance evaluations that these programs and accreditation bodies expect. Always confirm the specific testing and recordkeeping requirements with the authority having jurisdiction.

Frequently Asked Questions (FAQs)

What is the difference between direct and indirect conversion in a DR detector?

A direct-conversion detector uses a photoconductor such as amorphous selenium to turn absorbed X-rays straight into electric charge, collected under an applied field with almost no lateral spread. An indirect-conversion detector first turns X-rays into visible light in a scintillator such as cesium iodide or gadolinium oxysulfide, then converts that light to charge in an amorphous-silicon photodiode array. The extra optical step can capture the beam efficiently but introduces some light spreading.¹²⁶

Which detector type has better spatial resolution?

Direct-conversion a-Se generally shows a higher MTF because charge is collected with little lateral diffusion, so the presampled MTF approaches the pixel-aperture limit. Structured (columnar) CsI recovers much of that resolution by channeling light; unstructured phosphors blur finer detail more.⁵⁶

Which detector type is more dose efficient?

In general projection radiography, indirect CsI detectors have often demonstrated higher DQE at clinically relevant frequencies because of strong absorption in a thick columnar scintillator, though the best choice depends on the task and beam quality. Dose efficiency is quantified by DQE, not by conversion category alone.⁵⁶

Is computed radiography a direct or indirect detector?

CR uses a photostimulable storage phosphor read out later by a scanning laser. It is an indirect, light-mediated technology, but its offline readout and light spread generally place its image quality and dose efficiency below integrated flat-panel DR.²⁶

Does the detector type change patient dose?

The detector does not set patient dose by itself, but a more dose-efficient detector allows the same image quality at lower exposure, and a well-calibrated exposure index keeps techniques from drifting upward. Detector efficiency, exposure-index management, and protocol optimization together keep dose as low as reasonably achievable.⁶⁹

Key Takeaways

• DR detectors divide into direct conversion (a-Se photoconductor, X-ray to charge) and indirect conversion (scintillator to light to charge), and that single choice drives measurable image quality.¹²⁶
• Direct a-Se preserves spatial detail with minimal lateral spread, giving a high MTF near the pixel aperture; structured CsI recovers much of that sharpness while absorbing the beam efficiently.⁵⁶
• DQE ties image quality to dose; at zero frequency it is approximately the product of absorption efficiency and the Swank factor, which is why converter thickness and uniformity matter.⁵⁶
• For high-detail tasks such as mammography, direct a-Se's high-frequency performance is a strength; for general and chest radiography, indirect CsI's dose efficiency often leads.⁶⁸
• Exposure-index calibration and monitoring, following AAPM Task Group 116 and 151, are essential to prevent dose creep regardless of detector type.⁹¹¹
• Specify and accept detectors by measured MTF, NNPS, and DQE under IEC 62220-1-1 conditions, not by category label.⁵¹⁰

Conclusion

Direct and indirect conversion are two answers to the same problem — how to turn the X-ray shadow of a patient into a faithful digital image — and each answer carries measurable strengths. Direct a-Se trades a little absorption for outstanding sharpness and clean noise; indirect structured CsI trades a little high-frequency resolution for efficient, dose-sparing absorption. A physicist who reads a detector's MTF, NPS, and DQE, keeps its exposure index honest, and tunes QC to its known failure modes will get excellent, consistent images from either technology while keeping patient dose as low as reasonably achievable.⁵⁶⁹

How DRPS Can Help

Diagnostic Radiation Physics Services (DRPS) supports imaging facilities across Florida, Maryland, Virginia, Washington DC, California, Nevada, Pennsylvania, New York, New Jersey, and Delaware with acceptance testing, annual performance evaluation, and quality-control program design for digital radiography systems. Our board-certified medical physicists evaluate detector MTF, noise, DQE, and exposure-index calibration; build detector-specific QC that fits your equipment; and translate the physics of your detectors into practical protocol optimization. Explore our diagnostic radiography physics and medical physicist consulting services, or contact us to discuss your fleet.

Related Resources

• Detective quantum efficiency in digital radiography
• Flat-panel detector uniformity QC
• Digital radiography exposure index
• Computed radiography imaging plate QC
• Digital radiography lag and ghosting QC
• Diagnostic radiography physics services

References

1. Yaffe MJ, Rowlands JA. X-ray detectors for digital radiography. Phys Med Biol. 1997;42(1):1-39. doi:10.1088/0031-9155/42/1/001. doi.org
2. Rowlands JA. The physics of computed radiography. Phys Med Biol. 2002;47(23):R123-R166. doi:10.1088/0031-9155/47/23/201. doi.org
3. Chotas HG, Dobbins JT 3rd, Ravin CE. Principles of digital radiography with large-area, electronically readable detectors: a review of the basics. Radiology. 1999;210(3):595-599. doi:10.1148/radiology.210.3.r99mr15595. doi.org
4. Kotter E, Langer M. Digital radiography with large-area flat-panel detectors. Eur Radiol. 2002;12(10):2562-2570. doi:10.1007/s00330-002-1350-1. doi.org
5. Samei E, Flynn MJ. An experimental comparison of detector performance for direct and indirect digital radiography systems. Med Phys. 2003;30(4):608-622. doi:10.1118/1.1561285. doi.org
6. Cowen AR, Kengyelics SM, Davies AG. Solid-state, flat-panel, digital radiography detectors and their physical imaging characteristics. Clin Radiol. 2008;63(5):487-498. doi:10.1016/j.crad.2007.10.014. doi.org
7. Zhao W, Rowlands JA. X-ray imaging using amorphous selenium: feasibility of a flat panel self-scanned detector for digital radiology. Med Phys. 1995;22(10):1595-1604. doi:10.1118/1.597628. doi.org
8. Fahrig R, Rowlands JA, Yaffe MJ. X-ray imaging with amorphous selenium: optimal spectra for digital mammography. Med Phys. 1996;23(4):557-567. doi:10.1118/1.597688. doi.org
9. Shepard SJ, Wang J, Flynn M, et al. An exposure indicator for digital radiography: AAPM Task Group 116 (executive summary). Med Phys. 2009;36(7):2898-2914. doi:10.1118/1.3121505. doi.org
10. International Electrotechnical Commission. Medical Electrical Equipment — Characteristics of Digital X-Ray Imaging Devices — Part 1-1: Determination of the Detective Quantum Efficiency — Detectors Used in Radiographic Imaging. IEC 62220-1-1:2015. Geneva: IEC; 2015. webstore.iec.ch
11. Jones AK, Heintz P, Geiser W, et al. Ongoing quality control in digital radiography: report of AAPM Imaging Physics Committee Task Group 151. Med Phys. 2015;42(11):6658-6670. doi:10.1118/1.4932623. doi.org
12. U.S. Food and Drug Administration. Performance Standards for Ionizing Radiation Emitting Products: Diagnostic X-Ray Systems and Their Major Components. 21 CFR 1020.31. ecfr.gov