Time-of-Flight PET: How TOF Improves SNR

Time-of-Flight (TOF) is a PET reconstruction technology that measures the tiny difference in arrival time between the two annihilation photons to localize where along the line of response the event occurred—reducing image noise and improving quantitative accuracy. On a modern digital scanner the timing measurement is precise enough to confine each event to a few centimeters of the line of response rather than its full length, and that single physical constraint is what drives better contrast, more reliable standardized uptake values (SUVs), and faster or lower-dose scans.

In this edition of the PhysicsPulse^TM Series, we examine how TOF works, quantify the gain with the same equations physicists use at acceptance testing, benchmark the timing resolution of current detector generations, and explain how facilities across Florida, Maryland, Virginia, Washington DC, California, and Nevada can verify and protect that performance over the life of a TOF-enabled PET/CT system.

---

Introduction

TOF is no longer an exotic feature—it is the default on every current clinical PET/CT platform, and its benefit is now large enough to change protocols, not just images. Understanding the physics helps technologists, physicists, and administrators optimize protocols, defend SUV quantification, and justify capital decisions.

The reason TOF matters more today than it did a decade ago is detector technology. Early TOF systems built on photomultiplier tubes (PMTs) achieved coincidence timing resolution near 500–600 picoseconds (ps)¹. The current generation of silicon photomultiplier (SiPM)–based digital scanners has roughly halved that figure, reaching the low-200 ps range⁶, which doubles the spatial localization and roughly doubles the effective sensitivity gain. That progression—from a marginal advantage to a defining one—is why TOF performance now belongs on the acceptance-testing checklist alongside spatial resolution and sensitivity.

This guide walks through the underlying physics, the quantitative gain equations, a detector-generation benchmark table, a worked localization example, the clinical impact, optimization and QC practices, and the accreditation and regulatory context that surrounds a TOF PET program.

---

Topic Explanation: Time-of-Flight (TOF) in PET Imaging

TOF PET improves image quality, quantitative accuracy, and scan efficiency by enhancing how precisely annihilation events are localized along the line of response. Conventional PET knows only the line a photon pair traveled; TOF adds where on that line the annihilation most likely occurred.

Major PET vendors—including Siemens Healthineers, GE HealthCare, Philips Healthcare, and United Imaging Healthcare—offer advanced TOF-enabled PET/CT systems using silicon photomultiplier (SiPM) detectors and fast lutetium-based scintillators, achieving timing resolutions typically below 400 picoseconds.

A few terms recur throughout this guide and are worth defining up front:

• Line of response (LOR): the straight line connecting the two detector elements that registered a coincident 511 keV photon pair.
• Coincidence timing resolution (CTR): the full-width-at-half-maximum (FWHM) uncertainty in the measured arrival-time difference between the two photons, expressed in picoseconds. This is the single performance number that governs the size of the TOF gain.
• TOF localization window (Δx): the spatial uncertainty along the LOR that results from the timing resolution; smaller is better.
• Effective sensitivity gain: the factor by which TOF improves the noise properties of the reconstructed image for a fixed number of true coincidences, equivalent to scanning a non-TOF system with proportionally more counts.

For a complementary look at the radionuclides imaged on these systems, see Understanding Common Isotopes in PET & Radiopharmaceutical Therapy.

---

How Time-of-Flight Works

TOF estimates the annihilation position by measuring the arrival-time difference between the two 511 keV photons and converting it to a distance along the line of response. Here is the underlying physics.

In conventional PET imaging, when a positron annihilates with an electron, two 511 keV photons are emitted in (nearly) opposite directions. These photons define a Line of Response (LOR), but conventional PET cannot determine exactly where along that line the annihilation occurred. Reconstruction must therefore distribute the event uniformly along the full LOR, which adds noise.

TOF PET measures the arrival time difference between the two photons. Because photons travel at the speed of light, even small timing differences map to a position along the LOR. The position uncertainty is set by the system's coincidence timing resolution:

$$\Delta x=\frac{c\cdot \Delta t}{2}$$

where $\Delta x$ is the spatial uncertainty along the LOR, $c$ is the speed of light ($\approx 3\times 10^8$ m/s), and $\Delta t$ is the coincidence timing resolution. For example, a timing resolution of ~214 picoseconds corresponds to a localization window of roughly 3 cm along the LOR, confining the event to a segment of the line rather than its full length.

The factor of two appears because the difference in path length traveled by the two photons is twice the offset of the annihilation point from the LOR midpoint—each picosecond of timing error contributes to both legs of the coincidence. This is the foundational equation of TOF PET, and it is worth committing to memory because every downstream benefit traces back to it.

This results in:

• Improved localization accuracy along the LOR
• Reduced positional uncertainty during back-projection
• Enhanced signal localization, which suppresses noise in reconstruction

Worked Example: From Timing Resolution to Localization Window

A concrete calculation makes the relationship tangible. Suppose a digital SiPM scanner is specified at a coincidence timing resolution of $\Delta t=214\text{ps}=214\times 10^{-12}\text{s}$, consistent with the published Biograph Vision performance evaluation⁶. Substituting into the localization equation:

$$\Delta x=\frac{(3\times 10^8\text{m/s})(214\times 10^{-12}\text{s})}{2}=\frac{6.42\times 10^{-2}\text{m}}{2}\approx 3.2\text{cm}$$

So a 214 ps system localizes each annihilation to within about 3.2 cm along the LOR. By contrast, a legacy PMT-based TOF system at $\Delta t=550\text{ps}$ yields:

$$\Delta x=\frac{(3\times 10^8)(550\times 10^{-12})}{2}\approx 8.3\text{cm}$$

The digital system therefore confines each event to roughly 40% of the spatial span of the older system. Because the noise benefit of TOF scales with how small $\Delta x$ is relative to the patient, halving the localization window has a direct and measurable effect on image quality, as the next section quantifies.

The TOF Sensitivity Gain

The effective sensitivity gain from TOF scales with patient size. A commonly cited approximation is:

$$\text{Gain}\approx \frac{D}{\Delta x}$$

where $D$ is the diameter of the imaged object and $\Delta x$ is the TOF localization window. Larger patients—where conventional PET noise is worst due to greater attenuation and scatter—benefit most, which is why TOF is so valuable in the bariatric and large-torso population.

This variance-reduction factor $g\approx D/\Delta x$ describes the improvement in the noise variance of the reconstructed image: a TOF system behaves like a non-TOF system collecting $g$ times as many counts. Because image signal-to-noise ratio (SNR) improves with the square root of the count statistics, the corresponding SNR gain is²⁷:

$$\frac{\mathrm{SN}{\mathrm{R}}_{\mathrm{TOF}}}{\mathrm{SN}{\mathrm{R}}_{\mathrm{nonTOF}}}\approx \sqrt{\frac{D}{\Delta x}}$$

This square-root relationship between SNR gain, object size, and timing resolution is the central quantitative result of TOF PET, first established analytically and later confirmed in phantom and patient studies²⁷⁸.

Worked Example: SNR Gain for a Large Patient

Consider a large abdomen of effective diameter $D=40\text{cm}$ imaged on the 214 ps system above ($\Delta x\approx 3.2\text{cm}$). The variance-reduction gain is:

$$g\approx \frac{D}{\Delta x}=\frac{40\text{cm}}{3.2\text{cm}}\approx 12.5$$

That is, the TOF reconstruction has noise variance comparable to a non-TOF scan with roughly 12.5 times the counts. The corresponding SNR gain is:

$$\frac{\mathrm{SN}{\mathrm{R}}_{\mathrm{TOF}}}{\mathrm{SN}{\mathrm{R}}_{\mathrm{nonTOF}}}\approx \sqrt{12.5}\approx 3.5$$

For comparison, the same patient on a 550 ps legacy system ($\Delta x\approx 8.3\text{cm}$) yields $g\approx 4.8$ and an SNR gain of about $2.2$. The digital system delivers a markedly larger benefit for exactly the patients in whom conventional PET struggles most. Note that these analytic values represent an idealized ceiling; the gain actually realized in iterative reconstruction depends on the activity distribution, count level, and location within the object²⁸.

---

Key Technical Principles: TOF Timing Resolution by Detector Generation

The benefit of TOF is set almost entirely by one number—coincidence timing resolution—and that number has improved by roughly a factor of three across detector generations. The table below summarizes how detector technology maps to timing resolution, the resulting localization window from $\Delta x=c\Delta t/2$, and the approximate effective sensitivity gain for a representative 40 cm patient ($g\approx D/\Delta x$, SNR gain $\approx \sqrt{g}$).

Two patterns stand out. First, the move from PMT to SiPM photosensors and from slower to faster lutetium-based scintillators is what unlocked the sub-300 ps regime; the change is a hardware story about light yield, photodetector timing jitter, and electronics⁴⁵. Second, the localization window and the gain track timing resolution exactly through the two equations above—there is no separate "TOF magic," only the consequences of $\Delta x=c\Delta t/2$.

A practical caveat: coincidence timing resolution is a measured, drift-prone quantity, not a fixed nameplate value. Published evaluations report it as a range that can vary with count rate—the Biograph Vision, for instance, was characterized at 210–215 ps across the relevant count-rate range under NEMA NU 2-2012/2018 testing⁶. This is precisely why ongoing QC of the timing chain matters: the gain in the table is only available if the detector and electronics are calibrated and stable.

---

Major Commercial Time-of-Flight PET Systems

TOF capability is now standard across the major clinical PET/CT platforms. The systems below represent current and recent commercial offerings.

Siemens Healthineers

TOF technology is implemented across the Siemens PET/CT product line, including:

• Biograph Vision PET/CT
• Biograph Vision Quadra PET/CT
• Biograph mCT PET/CT
• Biograph Horizon PET/CT

These systems use fast LSO crystals and SiPM detectors to achieve timing resolution as low as ~214 picoseconds in the Vision platform⁶. For a related Siemens workflow technology, see Siemens PET Flow Technology: Continuous Motion PET.

GE HealthCare

GE incorporates TOF capability across multiple platforms, including:

• Discovery MI PET/CT series
• Discovery IQ PET/CT
• Discovery 690 / 710 PET/CT
• SIGNA PET/MR AIR

GE systems use SiPM detectors and advanced reconstruction algorithms to improve sensitivity, image quality, and quantitative accuracy.

Philips Healthcare

Philips uses its TruFlight TOF technology in systems including:

• Vereos Digital PET/CT
• Gemini TF PET/CT
• Gemini TF Big Bore PET/CT

Philips Vereos was the first fully digital clinical PET/CT system, using digital SiPM technology for improved timing performance.

United Imaging Healthcare

United Imaging offers TOF-enabled systems including:

• uMI 550 PET/CT
• uMI 780 PET/CT
• uEXPLORER total-body PET/CT

These systems emphasize high sensitivity, stable detector performance, and improved quantitative imaging. The uEXPLORER total-body platform, with its 194 cm axial field of view, was characterized under NEMA NU 2-2018 and demonstrates how a long axial field of view combines with TOF to produce ultra-high sensitivity⁹.

---

Key Technical Principles: Advantages of TOF PET

TOF delivers three coupled benefits: better image quality, better quantitative accuracy, and better scan efficiency—all stemming from the same reduction in positional uncertainty.

Improved Image Quality

TOF improves signal localization, which leads to:

• Higher lesion contrast
• Reduced image noise
• Improved visualization of small lesions

These improvements are especially beneficial in larger patients and low-count imaging conditions. Phantom and human-observer studies have repeatedly shown that adding TOF improves the area under the lesion-detection curve, with the largest gains in high-body-mass patients and low-count tasks³⁸.

Improved Quantitative Accuracy

TOF improves the accuracy of SUV (standardized uptake value) measurements and quantitative imaging by reducing statistical noise propagation during reconstruction. This results in:

• More reliable SUV values
• Improved lesion detectability
• Enhanced therapy response assessment

Reliable, reproducible SUVs are essential for serial response assessment and for theranostic dosimetry, where quantification feeds directly into treatment decisions. Because TOF also accelerates the convergence of contrast in iterative reconstruction, it tends to recover lesion uptake more completely at clinically practical iteration counts than non-TOF reconstruction⁵.

Improved Scan Efficiency and Dose Optimization

Because TOF improves effective sensitivity, facilities can achieve diagnostic image quality with:

• Shorter scan times
• Lower administered activity
• Improved patient throughput

This provides flexibility to optimize protocols based on clinical needs and supports ALARA-driven dose reduction. The same effective-sensitivity gain that improves image quality at fixed scan time can instead be "spent" on a shorter acquisition or a lower injected activity, and the clinical literature frames this explicitly as a trade among image quality, scan time, and dose²⁷. Shorter uptake-room and table times also have a workflow benefit; for how uptake timing itself affects quantification, see PET Uptake Time: Why the Interval Between Injection and Imaging Matters.

---

Clinical Impact

TOF translates directly into better diagnostic confidence and more efficient operations across oncology, cardiology, and theranostics. The sharper localization improves small-lesion detection in oncologic staging, the more reliable SUVs strengthen therapy-response monitoring, and the sensitivity gain enables either faster scans (higher throughput) or lower injected activity (lower patient and staff dose).

For larger patients, TOF can be the difference between a non-diagnostic study and a confident read. As the worked example showed, the SNR gain for a 40 cm abdomen is roughly 3.5-fold on a digital system, and the supporting equations explain why the benefit concentrates in exactly the patients where conventional PET is noisiest: a larger $D$ means a larger $D/\Delta x$. Human-observer studies confirm that the TOF advantage in lesion detection is greatest at high body mass index³.

For busy departments, the throughput gain compounds across a daily schedule. And for theranostic programs imaging agents tied to therapy—where quantification drives dosing—the quantitative stability TOF provides is a genuine clinical asset.

---

Practical Tips for Technologists and Physicists

To realize TOF's benefits in practice, the timing chain must be calibrated and stable, and protocols must be matched to the system's capabilities. Recommended practices:

• Complete daily QC, including timing calibration and coincidence timing resolution checks per the manufacturer's schedule.
• Follow manufacturer-recommended TOF reconstruction and acquisition protocols rather than disabling TOF or reverting to legacy presets.
• Keep patient positioning accurate and consistent, and minimize patient motion, which degrades both spatial and timing performance.
• Monitor coincidence timing resolution over time as a detector-health indicator; drift can erode the TOF gain before it is otherwise obvious.
• Validate SUV reproducibility with a NEMA-style phantom when changing reconstruction settings, software, or after major service.
• Treat coincidence timing resolution as a trend, not a single number: a 20–30 ps creep upward translates—through $\Delta x=c\Delta t/2$—into a measurable loss of localization and effective sensitivity, so log it and watch the trajectory.

Proper protocol selection lets facilities balance image quality, scan time, and radiation dose. For protocol-optimization principles that carry over from the CT side of PET/CT, see CT Protocol Optimization — Balancing Dose, Image Quality, and Compliance.

---

Regulatory Considerations

TOF PET/CT systems fall under the same accreditation, acceptance-testing, and radioactive-material regulations as any clinical PET program—TOF does not change the compliance obligations, but it raises the bar on quantitative validation. Key references:

• NEMA NU 2 defines the standardized performance measurements—including coincidence timing resolution and sensitivity—used to characterize TOF PET systems and to compare platforms objectively. The published performance evaluations cited throughout this article were all conducted under NEMA NU 2-2012 or NU 2-2018 protocols, which is what makes their timing-resolution figures comparable across vendors⁶⁹.
• ACR Nuclear Medicine and PET Accreditation requires phantom imaging, qualified medical physicist oversight, and documented QC; accredited PET imaging is also tied to reimbursement and to many state requirements.
• AAPM TG-126 provides guidance relevant to PET/CT acceptance testing and ongoing quality control.
• NRC 10 CFR 35 (or Agreement State equivalents) governs the medical use of the PET radiopharmaceuticals these systems image, including authorized-user and dosimetry requirements, while 10 CFR 20 sets the underlying occupational and public dose limits that ALARA-driven dose reduction supports. See our NRC Radioactive Material License Guide.
• State radiation control programs apply to facility registration, surveys, and shielding. Florida facilities, for instance, operate under 64E-5; see Florida Radiation Safety Requirements for Imaging Centers. Maryland, Virginia, Washington DC, California, and Nevada each have their own programs that DRPS supports. (Washington DC is regulated directly by the NRC; FL, MD, VA, CA, and NV are Agreement States.)

Because TOF systems are typically sited in PET/CT suites, their physical installation also raises shielding questions for both the 511 keV annihilation photons and the CT subsystem; for the design methodology, see our PET/CT Shielding Calculations Guide.

---

Frequently Asked Questions (FAQs)

How does TOF improve PET image quality?

TOF localizes annihilation events more precisely along the line of response by measuring the photon arrival-time difference. Confining each event to a short segment of the LOR—rather than its full length—reduces noise and improves contrast and quantitative accuracy.

What timing resolution do modern TOF PET scanners achieve?

Modern SiPM-based TOF PET systems typically achieve coincidence timing resolution below 400 picoseconds. The Siemens Biograph Vision platform reaches as low as ~214 picoseconds, corresponding to a localization window of roughly 3 cm along the LOR⁶.

How is TOF spatial uncertainty calculated from timing resolution?

The spatial uncertainty along the line of response is $\Delta x=c\Delta t/2$, where $c$ is the speed of light and $\Delta t$ is the coincidence timing resolution. A 214 ps system gives $\Delta x\approx 3.2$ cm; a 550 ps system gives about 8.3 cm. The factor of two arises because the path-length difference between the two photons is twice the offset of the annihilation point from the LOR midpoint.

How large is the TOF SNR gain?

The signal-to-noise ratio gain scales as the square root of the object diameter divided by the TOF localization window: $\mathrm{SN}{\mathrm{R}}_{\mathrm{TOF}}/\mathrm{SN}{\mathrm{R}}_{\mathrm{nonTOF}}\approx \sqrt{D/\Delta x}$. For a 40 cm patient on a 214 ps system ($\Delta x\approx 3.2$ cm), the variance-reduction gain $g\approx D/\Delta x$ is about 12.5 and the SNR gain is about 3.5. These are idealized analytic values; the realized gain in iterative reconstruction is count- and object-dependent²⁸.

Does TOF PET reduce patient radiation dose?

Indirectly, yes. Because TOF improves effective sensitivity, facilities can maintain diagnostic image quality with shorter scan times or lower administered activity. This supports ALARA and dose optimization without sacrificing diagnostic confidence.

Why does TOF benefit larger patients the most?

The TOF sensitivity gain scales with the diameter of the imaged object divided by the TOF localization window. Larger patients have longer lines of response and more attenuation and scatter—exactly where conventional PET noise is worst—so TOF removes proportionally more positional uncertainty.

What scintillators and detectors enable TOF PET?

TOF PET relies on fast lutetium-based scintillators such as LSO and LYSO paired with silicon photomultiplier (SiPM) detectors. Together they provide the rapid scintillation light and fast electronic timing needed to resolve sub-nanosecond arrival-time differences.

---

Key Takeaways

• TOF measures the photon arrival-time difference between the two 511 keV annihilation photons to localize the event along the line of response, reducing reconstruction noise.
• The localization window follows $\Delta x=c\Delta t/2$: a 214 ps system confines each event to ~3.2 cm, versus ~8.3 cm for a 550 ps legacy system.
• The SNR gain scales as $\sqrt{D/\Delta x}$, so a 40 cm patient on a digital system sees roughly a 3.5-fold SNR gain (a ~12.5-fold variance-reduction gain).
• Modern SiPM-based TOF PET achieves timing resolution below ~400 ps, with the Siemens Biograph Vision platform reaching ~214 ps⁶.
• The TOF gain scales with patient size, making it most valuable for large and bariatric patients where conventional PET is noisiest³.
• TOF improves SUV accuracy and reproducibility, strengthening therapy-response assessment and theranostic dosimetry.
• TOF enables shorter scans or lower administered activity at equal image quality, supporting throughput and ALARA-based dose reduction.
• TOF is standard, not optional, across current Siemens, GE, Philips, and United Imaging clinical PET/CT platforms.

---

How DRPS Can Help

Diagnostic Radiation Physics Services provides qualified medical physicist support for PET/CT programs across Florida, Maryland, Virginia, Washington DC, California, and Nevada. We perform PET/CT acceptance testing and annual surveys, NEMA-based performance verification (including coincidence timing resolution and sensitivity), SUV quantification validation, ACR accreditation support, and shielding evaluations—keeping your TOF systems performing to spec and your program compliant with NRC, Agreement State, and state-specific regulations.

If you are commissioning a new digital TOF platform, benchmarking timing resolution against the manufacturer's specification, or troubleshooting SUV drift, our board-certified medical physicists can help. Contact DRPS or review our service areas to get started.

---

Conclusion

Time-of-Flight is a major advancement in PET imaging that improves spatial localization, quantitative accuracy, and imaging efficiency by exploiting the arrival-time difference of annihilation photons. Modern TOF-enabled PET systems from Siemens, GE, Philips, and United Imaging are now standard in clinical practice and are central to delivering high-quality, quantitative PET imaging.

The physics is unusually clean: every benefit flows from the two equations $\Delta x=c\Delta t/2$ and $\mathrm{SN}{\mathrm{R}}_{\mathrm{TOF}}/\mathrm{SN}{\mathrm{R}}_{\mathrm{nonTOF}}\approx \sqrt{D/\Delta x}$. That is also why the right way to protect the benefit is to measure and trend coincidence timing resolution over the life of the system. Understanding TOF helps technologists and physicists optimize imaging protocols, improve diagnostic quality, and support the expanding role of PET in oncology, cardiology, and theranostics.

---

Related Resources

• Common PET & RPT isotopes
• Siemens PET Flow (continuous bed motion)
• PET uptake time and quantification
• CT protocol optimization
• NRC radioactive material license guide
• PET/CT and nuclear medicine physics

References

1. Karp JS, Surti S, Daube-Witherspoon ME, Muehllehner G. Benefit of time-of-flight in PET: experimental and clinical results. J Nucl Med. 2008;49(3):462-470. doi:10.2967/jnumed.107.044834. doi.org
2. Conti M. Focus on time-of-flight PET: the benefits of improved time resolution. Eur J Nucl Med Mol Imaging. 2011;38(6):1147-1157. doi:10.1007/s00259-010-1711-y. doi.org
3. Surti S, Scheuermann J, El Fakhri G, et al. Impact of time-of-flight PET on whole-body oncologic studies: a human observer lesion detection and localization study. J Nucl Med. 2011;52(5):712-719. doi:10.2967/jnumed.110.086678. doi.org
4. Surti S, Karp JS. Advances in time-of-flight PET. Phys Med. 2016;32(1):12-22. doi:10.1016/j.ejmp.2015.12.007. doi.org
5. Vandenberghe S, Mikhaylova E, D'Hoe E, Mollet P, Karp JS. Recent developments in time-of-flight PET. EJNMMI Phys. 2016;3(1):3. doi:10.1186/s40658-016-0138-3. doi.org
6. van Sluis J, de Jong J, Schaar J, et al. Performance characteristics of the digital Biograph Vision PET/CT system. J Nucl Med. 2019;60(7):1031-1036. doi:10.2967/jnumed.118.215418. doi.org
7. Surti S. Update on time-of-flight PET imaging. J Nucl Med. 2015;56(1):98-105. doi:10.2967/jnumed.114.145029. doi.org
8. Surti S, Karp JS, Popescu LM, Daube-Witherspoon ME, Werner M. Investigation of time-of-flight benefit for fully 3-D PET. IEEE Trans Med Imaging. 2006;25(5):529-538. doi:10.1109/TMI.2006.871419. doi.org
9. Spencer BA, Berg E, Schmall JP, et al. Performance evaluation of the uEXPLORER total-body PET/CT scanner based on NEMA NU 2-2018 with additional tests to characterize PET scanners with a long axial field of view. J Nucl Med. 2021;62(6):861-870. doi:10.2967/jnumed.120.250597. doi.org
10. National Electrical Manufacturers Association. NU 2: Performance Measurements of Positron Emission Tomographs (PET). Rosslyn, VA: NEMA. nema.org
11. Conti M. State of the art and challenges of time-of-flight PET. Phys Med. 2009;25(1):1-11. doi:10.1016/j.ejmp.2008.10.001. doi.org
12. Cherry SR, Sorenson JA, Phelps ME. Physics in Nuclear Medicine. 4th ed. Philadelphia, PA: Elsevier Saunders; 2012. elsevierhealth.com
13. Siemens Healthineers. Biograph Vision PET/CT Technical Specifications. siemens-healthineers.com
14. GE HealthCare. Discovery MI PET/CT System Specifications. gehealthcare.com
15. Philips Healthcare. Vereos Digital PET/CT Technical Overview. philips.be
16. United Imaging Healthcare. uMI PET/CT System Specifications. united-imaging.com

---

PhysicsPulse^TM Series

Troy Zhou, PhD, DABR, DABSNM

Diagnostic Radiation Physics Services