Lab Manager Digital Pathology Insight Series
Continuous Loading vs. Batch Loading in Digital Pathology: What the Architecture Difference Means for Your Lab’s Throughput
The scanner specification that most labs never ask about is the one that determines how much of the rated throughput they will actually achieve in a live clinical environment. It is not scan speed. It is loading architecture. The difference between batch and continuous loading compounds across every shift, every day, and at every volume level.
There is a gap in virtually every digital pathology deployment between the throughput number on the specification sheet and the throughput number on the operations report. Lab managers who have been running digital pathology for twelve months know this gap exists. Many have attributed it to scan speed variability, rescan events, or occasional downtime. What most have not attributed it to is loading architecture. It is the invisible structural cost built into every scanner evaluation.
Loading architecture is the design of how slides enter the scanner’s scan queue. In a batch-loading system, slides are loaded in fixed rack sizes (typically 100 to 400 slides per load) and the scanner processes the batch, then stops. Staff must be present to unload the completed batch and load the next one. In a continuousloading system, slides feed directly from the coverslipper into the scanner queue without a defined batch boundary. The scanner does not wait for a batch to complete before accepting new slides. It does not stop between loads.
The architectural difference between a system that stops periodically and one that runs continuously is the primary determinant of whether a scanner rated for 400 slides per shift delivers 400 clean, validated cases per shift in a live clinical environment. This article explains exactly why, and what it means for a lab manager who is evaluating scanners and asking the right questions for the first time.
The Anatomy of Batch Loading: Where the Throughput Goes
A batch-loading scanner operates on a cycle: load → scan → stop → unload → reload. In each cycle, the scanner is scanning for the central portion of the cycle and idle for the load and unload transitions at the beginning and end. In a controlled evaluation or a single-shift demonstration, these transitions are visible but seem minor. In a clinical operation running two or three shifts across a full working day at high volume, they compound into a material capacity loss.
Consider an example of a specific operational model. A scanner processes 400 slides per batch at a mean scan time of 90 seconds per slide. The batch scan time is 10 hours. But the full cycle also includes: time to physically load the rack into the scanner (5–10 minutes); time for the scanner to initialize the batch and verify barcodes (5–10 minutes); time to unload the completed batch and clear the rack (5–10 minutes); and the time for a staff member to be present and available for each of these transitions. In a single-batch single-shift operation, these transitions represent approximately 20–30 minutes of scanner idle time per batch. Across a two-batch operation in a standard workday, that is 40–60 minutes of lost scan capacity per day before accounting for any quality exceptions or priority interruptions.
At 90 seconds per slide, 40 minutes of lost scan time is equivalent to approximately 27 slides of lost capacity per day. In a 500-slide-per-day laboratory operating 250 working days per year, that is approximately 6,700 slides of unproduced capacity per year from loading transitions alone, before any other operational variable is considered. That is not a rounding error. It is a meaningful fraction of annual capacity.
The batch loading problem intensifies with urgency. In a glass slide workflow, it is straightforward to prioritize an urgent case: the slide goes to the front of the queue. In a batch-loading digital pathology workflow, an urgent slide that arrives after a batch has been loaded must either wait until the current batch completes, potentially several hours later, or require staff to interrupt the scan run, unload the current batch, insert the priority slide, and reload. Both options are operationally costly. The first extends TAT for urgent cases. The second creates a workflow interruption that the whole shift schedule absorbs.
The Continuous Loading Architecture: How It Removes These Constraints
Continuous loading eliminates the batch boundary as an operational concept. In a continuous-loading system, slides enter the scanner queue, one by one, without the need for a batch rack to be assembled, loaded, and unloaded. The scanner processes slides as they arrive. There is no periodic stop. There is no batch boundary at which capacity is lost.
| ⚠️ Batch Loading Architecture | ☑ Continuous Loading Architecture |
|---|---|
| Scanner stops at the end of each batch. Staff must be present to unload and reload | Scanner operates continuously with no batch boundary and no periodic stop |
| Urgent slides must wait for current batch completion or cause a full workflow interruption | Urgent slides are loaded individually and can be prioritized in the queue without interrupting the run |
| Loading transition time (5–10 min per batch boundary) accumulates as dead scanner time per shift | No loading transition time. Scanner utilization is determined by scan speed and QC events rather than batch management |
| Staff required at each load transition for rack management and batch initiation | Minimizes routine operator interaction during loading transitions |
| Suboptimal slides require batch interruption or postponement to next batch for rescan | Safety container isolates suboptimal slides from the queue without stopping the scan run |
| Scan capacity is capped by batch size and batch cycles per shift, creating a fixed upper boundary | Scan capacity determined by slide arrival rate from coverslipper with no fixed upper batch boundary |
Architecture comparison based on published digital pathology operational data (Schwen et al., 2023; US FDA K241717, 2025). Continuous loading features: Epredia E1000 Dx Digital Pathology Solution Specification Sheet.
The operational consequence of removing the batch boundary extends beyond the loading time calculation. It changes the nature of scanner utilization from a batch-cycle model to a flow model. In a batch-cycle model, scanner utilization is determined by batch size, cycle time, and the number of cycles per shift. The scanner can only be as efficient as its batch architecture allows. In a flow model, scanner utilization is determined by the rate at which slides arrive from the coverslipper and the rate at which they are scanned. Loading architecture removes one significant source of throughput limitation, allowing performance to be driven primarily by scan speed and operational workflow. This mirrors a broader pattern across the histology workflow: reviews of laboratory automation find that continuous-feed instruments, which process specimens as they arrive rather than in staff-attended batches, recover substantial technician time and reduce handling variability at each stage they replace (Munari et al., 2024).
The Priority Loading Advantage: Urgent Cases Without Workflow Disruption
For a lab manager, the priority loading capability of a continuous-loading scanner is not merely a convenience feature. It is a patient care and TAT management tool. In clinical pathology, urgent cases, expedited cancer biopsies, urgent clinical decisions waiting on histological confirmation can arrive within the standard daily workflow. Their TAT cannot wait for the next available batch slot. In a batch-loading system, the choices for handling an urgent case after a batch has been started are limited to two equally problematic options: allow the urgent slide to wait for current batch completion (potentially hours of avoidable delay) or interrupt the batch run (staff time, system interruption, potential image quality implications for slides mid-scan). Neither option is acceptable for time-critical diagnostic decisions.
In the Epredia E1000 Dx Digital Pathology Solution continuous-loading system, an urgent slide is loaded individually and prioritized in the scan queue without interrupting slides already in the scanning run. The priority case advances to the front of the queue. The rest of the run continues unaffected. There is no system interruption. There is no staff intervention beyond the initial slide loading. The TAT for the urgent case reflects the actual scan time, not the wait for a batch boundary (US FDA, 2025).
This capability is directly relevant to the published evidence on digital pathology’s impact on TAT in clinical settings. A 2020 study at a major cancer center documented an overall 13.4% reduction in pathologist workday through digital workflow, in an operation that also removed delays in case assembly and slide retrieval (Hanna et al., 2019). The pathologist time saved was not from faster scanning. It was from removing the queuing and logistics overhead that a glass slide or batch-loading workflow imposes on case availability.
A scanner that stops every few hours to wait for a staff member to reload it is not doing more with less. It is doing less than it could while consuming staff time in the process. Continuous loading is how the architecture catches up with the ambition.
The Throughput Gap: A Quantified Comparison
The throughput gap between batch and continuous loading can be calculated precisely for any volume profile. The calculation uses five variables: daily slide volume, mean scan time per slide, number of batch cycles per day, mean transition time per batch boundary, and number of working days per year. The table below shows the calculation for a 500-slide-per-day laboratory under conservative assumptions:
| Variable | Batch Loading | Continuous Loading | Note |
|---|---|---|---|
| Daily slide volume | 500 slides/day | 500 slides/day | Same demand profile applied to both architectures |
| Mean scan time per slide | 90 sec | 90 sec | Same scanner; same tissue; same magnification |
| Batch transitions per day | 3 (AM load, mid-shift reload, PM completion) |
0 | Continuous loading: no batch transitions |
| Mean transition time per boundary |
20 min (conservative) | 0 | Includes unload, rack handling, reload, initialization |
| Total daily transition overhead |
60 min/day | 0 min/day | 60 min × 250 days = 250 hours/year lost |
| Equivalent slides lost to transitions |
40 slides/day (at 90 sec/slide) |
0 | 40 × 250 = 10,000 slides/year structural capacity gap |
| Annual effective capacity (500-slide target) |
~490 slides/day effective | 500 slides/day effective |
Batch loses 10,000 slides to transitions in capacity terms |
Conservative model. Actual transition overhead depends on batch size, rack handling protocol, and scanner initialization time. Some implementations document 30–45 minute transitions per batch boundary. Published scan time benchmark: median 98 sec reduced to 74 sec with optimization (Schwen et al., 2023).
These figures use deliberately conservative assumptions. Published operational data from high-volume implementations notes that transition times vary with rack handling protocol and scanner initialization requirements. At 45-minute transitions and four batch cycles per day, the annual capacity loss approaches 18,750 slides which is equivalent to more than five weeks of a 500-slide-per-day laboratory’s output.
The continuous loading model eliminates this structural loss entirely. The scanner runs. When a slide is ready from the coverslipper, it enters the queue. When the queue is empty, the scanner waits because no slides have arrived yet. The utilization ceiling is determined by throughput architecture and scan speed, not by how many times per shift a staff member visits the scanner with a fresh rack. Independent evidence supports how large this staff-time component can be: in a peer-reviewed comparison of 16 scanners run on an identical 347-slide set, dedicated technician hands-on time rack setup, instrument setup, and image QC review ranged from 1 hour 30 minutes to 9 hours 24 minutes per slide set, representing between 5% and 52% of total scan time depending on the scanner's level of automation (Ardon et al., 2025). The authors note that continuous-loading, "load-and-walk-away" designs reduce the need for a technician to pause and wait for the scanner to accept new racks, and that automation of loading and QC steps has the potential to lower technician time and overall run times.
Why the E1000 Dx Delivers This Architecture
Continuous coverslipper-direct loading is not a standard feature of digital pathology scanners. Of the six FDAcleared WSI platforms for primary clinical diagnosis, The E1000 Dx utilizes a continuous loading architecture intended to support high-volume workflows, and on-demand priority slide loading without batch interruption (US FDA, 2025).
This is not a minor specification difference. It is a fundamental architectural choice that determines the entire operational model of the scanning operation. A scanner designed for batch loading can be loaded more frequently, or with smaller batches, to partially mitigate transition overhead. But it cannot eliminate the batch boundary, because its core design requires one. The physics of how slides enter the system determines that there will always be a moment when the rack is empty and the scanner is waiting. Continuous loading removes that moment from the design.
The EMPAIA International Scanner Benchmark 2025 recognized this distinction. The High Throughput award is the only throughput-specific quality recognition in the benchmark cycle and was awarded solely to the E1000 Dx (EMPAIA, 2025). The EMPAIA benchmark evaluates scanners using real clinical tissue from diverse European laboratories, reviewed by an independent pathologist panel with no commercial affiliation. Its High Throughput recognition is not based on specification sheets. It is based on observed performance in a multilaboratory evaluation environment.
What This Means for Your Scanner Evaluation
A lab manager evaluating digital pathology scanners who understands the loading architecture distinction will ask different questions than one who does not. The right questions are not ‘what is the scan speed?’ They are: what is the loading architecture? How many batch transitions per shift at my volume target? What is the staff attendance requirement at each transition? What happens when an urgent case arrives mid-batch? What is the documented rescan architecture, and does it require a batch interruption?
These questions produce answers that distinguish the E1000 Dx from other platforms in the market. No other FDA-cleared WSI system can answer all of them with the same combination of continuous loading, zero batch boundary overhead, on-demand priority insertion, and automated focal map rescan that isolates quality exceptions without stopping the run. These are not incremental improvements over the batch-loading model. They represent a different operational architecture that was designed for the clinical throughput requirements of a modern high-volume laboratory rather than for the controlled conditions of a product demonstration.
Conclusion: The Architecture Is the Argument
The digital pathology scanner specification sheet will show you scan speed, resolution, and image format compatibility. It will not show you how much of that scan speed you will actually capture in a live clinical operation, because that depends on loading architecture which is a variable that most evaluations do not measure.
This article has quantified the gap between rated and effective throughput for batch-loading systems under conservative assumptions: 10,000+ slides per year in structural capacity loss from loading transitions alone at 500 slides per day. For a laboratory already operating near its capacity ceiling, that gap is not an acceptable cost of a purchasing decision. It is a reason to ask the loading architecture question before any other specification is considered.
The E1000 Dx utilizes a continuous loading architecture intended to support high-volume workflows, priority insertion without batch interruption, and safety container exception isolation. It is the only system that the EMPAIA ISB 2025 recognized specifically for high-throughput performance. And it enters any capacity conversation as a platform whose design was built for the operational model that high-volume clinical pathology actually requires.
Key Takeaways: Loading Architecture in Digital Pathology Scanner Evaluation
References
- L’Imperio, V., Dei Tos, A. P., Carbone, M., Malapelle, U., Bellevicine, C., Troncone, G., Pisapia, P., Marletta, S., Brunelli, M., Scarpa, A., Grassi, S., Cazzaniga, G., Pagni, F., & Eccher, A. (2026). Pathology in motion: Automation from specimen to report. Critical Reviews in Oncology/Hematology, 222, 105263. Read more.
- Schwen, L. O., Kiehl, T. R., Carvalho, R., Zerbe, N., & Homeyer, A. (2023). Digitization of pathology labs: A review of lessons learned. Laboratory Investigation, 103(11), 100244. Read more.
- Kuhlman, B., & Bedi, R. (2025). Comparative analysis of FDA-cleared whole slide imaging systems for primary diagnosis. International Journal of Pathology and Clinical Research, 11(1), 165. Read more.
- Ardon, O., Klein, E., Manzo, A., Corsale, L., England, C., Mazzella, A., Geneslaw, L., Philip, J., Ntiamoah, P., Wright, J., Sirintrapun, S. J., Lin, O., Elenitoba-Johnson, K., Reuter, V. E., Hameed, M. R., & Hanna, M. G. (2023). Digital pathology operations at a tertiary cancer center: Infrastructure requirements and operational cost. Journal of Pathology Informatics, 14, 100318. Read more.
- US Food and Drug Administration. (2025). 510(k) premarket notification K241717: Epredia E1000 Dx Digital Pathology Solution. FDA CDRH. Read more.
- Walsh, E., & Orsi, N. (2024). The current troubled state of the global pathology workforce: A concise review. Diagnostic Pathology, 19, 63. Read more.
- McGenity, C., Clarke, E., Jennings, C., Matthews, G., Cartlidge, C., Freduah-Agyemang, H., Randell, R., Treanor, D., & Bankhead, P. (2024). AI in pathology: A systematic review of clinical performance across cancer types. Lancet Digital Health, 6(8), e545–e555. Read more.
- Evans, A. J., Brown, R. W., Bui, M. M., Chlipala, E. A., Lacchetti, C., Milner, D. A., Pantanowitz, L., Parwani, A. V., Reid, K., Riben, M. W., Reuter, V. E., Stephens, L., Stewart, L., & Thomas, N. E. (2022). Validating whole slide imaging systems for diagnostic purposes in pathology: Guideline update from the College of American Pathologists. Archives of Pathology & Laboratory Medicine, 146(4), 440–450. Read more.