Flat panel detector resolution is specified in line pairs per millimeter (lp/mm), and the limiting value is governed almost entirely by the detector pixel pitch: a published 175 µm pitch panel reaches about 3 lp/mm limiting resolution [S6], while a 139 µm selenium TFT prototype delivered 40% MTF at 2 lp/mm and 16% DQE at 1.5 lp/mm [S1].
For OEM system builders, the lp/mm number is not a marketing headline but a derivative of three hardware choices: pixel pitch, scintillator (CsI needle vs GOS powder), and TFT fill factor. Radiological practice expresses the same metric as spatial frequency, where one line pair equals one dark line plus one bright line of equal width [S5], and the absence of it is referred to clinically as blur [S4].
How Lp/mm Maps to Pixel Pitch and Nyquist
Line pairs per millimeter and pixel pitch are the same physical quantity viewed from two sides: a 100 µm pixel pitch produces 1/(2 × 0.1 mm) = 5 lp/mm at the Nyquist limit, and a 175 µm pitch lands at roughly 2.86 lp/mm, which rounds to the 3 lp/mm cited for clinical flat panels [S6]. The relationship is mechanical, not approximate, and is the same formula used in industrial line-scan camera work, where line-scan camera sensors size pixels against the smallest feature the line is meant to resolve.
Two corollaries follow for spec writers. First, halving the pixel pitch from 175 µm to 87.5 µm roughly doubles the Nyquist limit, but only if the scintillator does not blur the optical signal before it reaches the photodiode [S3]. Second, lp/mm describes the cutoff, not the contrast: a system with 5 lp/mm Nyquist can still show 50% MTF at 2 lp/mm and almost zero contrast at 4 lp/mm, which is why detector MTF curves are always plotted against spatial frequency, never stated as a single number [S3].
Scintillator Choice: CsI vs GOS and the Resolution Trade-off
CsI (cesium iodide) needle scintillators channel visible light toward the photodiode and reduce lateral spread, while GOS (gadolinium oxysulfide) powder scintillators spread light more and cost resolution at a given pixel pitch [S3]. The practical effect is that two flat panels with identical 150 µm pixel pitch can show different effective lp/mm once the scintillator blur is included, with the CsI unit holding higher MTF into the 3-4 lp/mm range and the GOS unit falling off earlier.
Thickness compounds the trade: thicker scintillators stop more X-ray photons and improve dose efficiency, but the longer light path increases blur, so high-resolution mammography panels tend toward thinner CsI, while general radiography panels trade some lp/mm for dose efficiency [S3]. Super-resolution digital radiography protocols have demonstrated roughly 3.5 lp/mm at 12.5 frames per second with a 75% dose reduction versus standard acquisition, illustrating how frame rate, dose, and resolution interact [S2].
MTF, DQE, and What the Lp/mm Number Actually Means
Limiting lp/mm is the spatial frequency at which MTF drops to a few percent, while DQE captures how much of the input signal-to-noise ratio survives the imaging chain; the selenium prototype cited above held 40% MTF at 2 lp/mm and 16% DQE at 1.5 lp/mm, and these two numbers together are a more honest spec than the headline Nyquist figure [S1]. For an OEM buyer, asking only for lp/mm without the MTF at a specific frequency is the single most common spec error in flat panel procurement, and a 5 lp/mm panel with 10% MTF at 2 lp/mm is not equivalent to a 3 lp/mm panel with 50% MTF at the same frequency [S3].
The same principle applies in NDT work where digital radiography has largely displaced film for weld inspection, and a spec-driven comparison of the two is laid out in film vs digital radiography for weld inspection: both modalities express resolution in lp/mm, but the comparison has to be done at matched MTF, not at the limiting frequency alone.
Who Needs High Lp/mm and Who Does Not
Mammography, where microcalcifications as small as 0.1 mm drive clinical decisions, demands the highest lp/mm flat panels on the market, typically well above 5 lp/mm Nyquist [S3]. Orthopedic imaging of fine cortical bone, pediatric chest work, and high-magnification NDT of small features sit in the same high-resolution bracket, while general radiography, fluoroscopy, and large-area security screening usually run 2-3 lp/mm because the relevant features are larger and dose efficiency matters more than sharpness.
The wrong choice is also a procurement failure mode: a 100 µm pitch panel specified for a general-rad room adds cost, increases quantum noise per pixel, and can force a higher dose to the patient or the test piece, with no diagnostic gain because the clinical features of interest sit below the new resolution floor anyway [S3]. The right panel is the one whose pixel pitch and scintillator match the smallest feature the workflow actually has to resolve, not the highest lp/mm the budget will allow.
Calculator Tooling and Cross-Checking the Spec
An lp/mm to pixel count calculator lets a buyer enter sensor size, target line pair frequency, and a Nyquist sampling margin, then returns the required pixel count and pixel pitch; this is the same back-of-envelope math the Edmund Optics resolution note and the FF calculator expose, where 1 lp/mm equals one full dark-plus-light cycle per millimeter and Nyquist demands at least two samples per cycle [S5][S9].
For process engineers, the discipline is to start from the smallest feature that must be resolved in the workflow, convert it to lp/mm, and only then pick a panel whose Nyquist comfortably exceeds that figure, typically with a 1.5-2× sampling margin to keep MTF usable at the working frequency rather than at the cutoff [S3][S5]. Resolution and dose efficiency, not resolution alone, are the two numbers that have to be co-specified, and any flat panel quote that gives only lp/mm without MTF-at-frequency and DQE should be sent back for a second data sheet [S1][S3].
Track for the next update: revised IEC 62220-1 acceptance limits for DQE measurement on indirect flat panels, and any 2026 vendor datasheets that publish MTF at 2, 3, and 4 lp/mm rather than at Nyquist alone, since the latter has become the limiting factor in OEM flat panel selection [S1][S3].
For the relevant spec sheets and selection criteria, see flat belt, and molding line.