How long does a mass spectrometry imaging experiment take? What factors slow it down?

The time required for a single mass spectrometry imaging run ≈ number of scan points × acquisition time per point, ranging from a few minutes to tens of hours. The main factors contributing to delays include: excessively high spatial resolution, excessive scan area, excessively long integration time per point, slow analyser acquisition, the need for multiple sample runs, and the computational overhead of data reconstruction and calibration. Optimising the resolution and scan area is the most direct way to speed up the process.

Let us first examine the basic formula. The total number of points for imaging is (area ÷ step size²), and the total duration ≈ number of points × time per point. For example, for a 1 cm² area with a step size of 50 μm, there are approximately 40,000 points; if each point takes 0.1 seconds, the total time is approximately 4,000 seconds (over 1 hour). Halving the step size quadruples the number of points, and the time required also increases approximately fourfold.

Resolution is the key variable. The higher the resolution (the smaller the step size), the greater the number of points and the longer the duration. The New Ceyuan MSI LDPI (2–3 μm) is suitable for microstructures but involves a huge number of points, whilst the MSI DPI (20–200 μm) requires far fewer points; Selecting the resolution based on structural scale can save a considerable amount of time.

The acquisition time per point depends on the analyser. TOF acquires the full spectrum rapidly at a high pulse frequency; the ProC-2 has a sampling rate of 30,000 Hz, resulting in extremely short acquisition times per point. If an analyser requiring long integration times is used, the time per point increases exponentially, leading to a significant cumulative delay.

Area and repetition can also add to the time required. The larger the scan area, the longer it takes; some methods require multiple sections or multiple rounds of staining/derivatisation, resulting in even longer cumulative times. Sample instability may also force you to slow down or repeat the process, further extending the time required.

Do not forget the post-processing overhead. The storage of vast amounts of point spectra, coordinate alignment, image reconstruction and calibration also take time; although not counted as part of ‘acquisition’, they form part of the complete experimental cycle. By using a high-throughput TOF system, a reasonable resolution and an automated workflow, the end-to-end time can be kept within an acceptable range.

Is it possible to view the images whilst scanning?

Yes. Most imaging software supports real-time stitching previews, allowing you to see the distribution map gradually taking shape during the scan, which facilitates the early detection of parameter issues; however, the final clear image will still require waiting until all points have been scanned and calibrated.

Will increasing the sampling rate always result in a proportional increase in speed?

Increasing the sampling rate reduces the time per point; however, when the time per point is already significantly faster than the mechanical stepping and trigger overhead, the bottleneck shifts to the motion platform, at which point further increases in sampling rate yield diminishing returns.

How can one minimise the time taken for a single image acquisition?

Scan only the target area rather than the entire sample, select a resolution that is just sufficient, use high-sampling-rate TOF, reduce unnecessary pre-processing cycles, and complete the acquisition within the sample’s stability window.