A Practical Approach to Troubleshooting a Malfunctioning MLC

A multileaf collimator (MLC) is central to modern radiotherapy, shaping fields for three-dimensional conformal treatment, intensity-modulated radiotherapy (IMRT), and volumetric modulated arc therapy (VMAT). When a leaf-position warning, failed interlock, or unexplained dose discrepancy appears, the priority is to protect the patient while identifying whether the problem is mechanical, electronic, software-related, or dosimetric.

A practical response begins with restraint. Do not repeatedly reset the treatment unit, override an interlock, or continue treatment simply because the fault disappears temporarily. Preserve the error message, treatment record, machine state, and affected plan. These details can distinguish a transient communication issue from a recurring leaf-drive or position-monitoring fault.

The troubleshooting process should connect machine behaviour with clinical risk. A single static-field error may have a different significance from an MLC problem during a long modulated arc, where small positional deviations can accumulate across hundreds of control points. The intended treatment technique, dose distribution, field complexity, and patient anatomy all matter.

For Australian departments, the approach also needs to fit local practice. A metropolitan service in Sydney or Melbourne may have rapid access to vendor engineers and a second linac, while a regional centre may rely on remote support and carefully planned downtime. Documentation should align with the department’s quality system, ACPSEM expectations, local radiation regulator requirements, and the realities of the Australian radiotherapy equipment market.

Establish a Safe Clinical Boundary

Stop treatment when the MLC fault affects beam delivery, leaf position verification, field shaping, or the reliability of the treatment record. Confirm whether radiation was delivered, how much was delivered, and whether the beam terminated normally. The radiation therapist should secure the treatment area and record the treatment fraction, field or arc, gantry angle, monitor units, and displayed fault.

The medical physicist should then review the machine console, treatment management system, and record-and-verify system without altering the original evidence. Capture screenshots where permitted, retain log files, and note the exact time. A fault that occurs only at a particular gantry angle or during a specific arc may provide an important clue.

Check whether the issue is isolated to one patient plan or appears during a controlled test. Do not use a patient as the diagnostic test object. If treatment must be completed urgently, consider an approved alternative machine or a clinically reviewed backup technique rather than an informal workaround.

Classify the Failure Before Testing

First separate a hard interlock from a soft warning. A hard interlock normally prevents irradiation or terminates the beam, while a soft warning may indicate a tolerance breach, communication delay, or failed pre-treatment verification. The distinction should be confirmed from the manufacturer’s documentation and local procedures, not inferred from the wording alone.

Next determine whether the failure is static or dynamic. Static issues may involve a leaf failing to reach a commanded position, excessive backlash, a damaged leaf tip, or a position-sensor problem. Dynamic faults can emerge only while leaves travel, during gantry rotation, or when the controller must coordinate leaf motion with dose rate and gantry speed.

Useful initial classifications include:

• A single leaf or bank repeatedly failing at a similar position
• Several neighbouring leaves showing correlated errors
• A fault linked to gantry angle, dose rate, or arc direction
• A discrepancy between commanded, reported, and measured positions
• A plan-specific failure associated with high modulation or small segments

Look for patterns across logs and previous quality-control results. A recurring error in leaf 23 is more suggestive of a local mechanical or encoder issue than a random planning-system defect. Conversely, simultaneous errors across many leaves may indicate a controller, network, calibration, or power problem.

For advanced techniques, clinical context is important. A department reviewing a stereotactic lung plan can use the SBRT planning context to appreciate why a small geometric error may be consequential when margins are tight and dose gradients are steep.

Check Mechanics, Control, and Communication

Begin with visual inspection under the department’s approved service procedure. Look for an obstruction, debris, unusual leaf noise, collision evidence, or a leaf that does not return to its expected position. Do not place hands or tools inside the collimator assembly unless authorised isolation and service procedures are in place.

Run manufacturer-approved diagnostic patterns, such as open fields, leaf sweeps, picket-fence sequences, or controlled static positions. Compare the observed motion with the electronic position readout. Listen for irregular drive sounds, delayed motion, or a leaf that hesitates before catching up. These observations should be recorded rather than treated as proof of a root cause.

Review communication logs between the treatment control system, MLC controller, and record-and-verify system. Timeouts, dropped messages, clock mismatches, or corrupted plan transfers can resemble a mechanical problem. Confirm that the correct beam model, energy, jaw settings, leaf sequence, and treatment machine were selected.

Australian sites should also consider operational context. A linac that has undergone recent servicing, software patching, power interruption, or network work may need a broader systems review. Vendor escalation is appropriate when a fault returns after reset, appears across multiple plans, or involves a safety-critical monitor.

Quantify the Dosimetric Consequence

Once the machine is stable and testing is authorised, establish whether the failure affected dose delivery. Compare planned and delivered monitor units, leaf trajectories, jaw positions, dose rate, and gantry motion. Use the treatment log and independent measurement rather than relying solely on the console’s statement that the beam completed.

For an interrupted fraction, determine the delivered segment or control-point range and whether the record-and-verify system has stored an accurate delivery history. A repeat fraction should never be prescribed by simply replaying the original plan without clinical and physics review. The physician, radiation therapist, and physicist need a clear account of what was delivered and what remains outstanding.

Patient-specific QA can help assess risk, but it is not a substitute for investigating the machine fault. For IMRT or VMAT, compare gamma results, ion chamber readings, film or array data, and reconstructed dose where available. A passing result may still conceal a clinically relevant local error, particularly in small targets or steep dose gradients.

VMAT brings additional coordination between leaves, dose rate, and gantry speed; the VMAT implementation lessons provide useful context for understanding why delivery complexity should influence the level of review.

Restore Service Through Evidence

After repair or calibration, repeat the failed test and a representative set of baseline checks. Include the affected leaf or bank, neighbouring leaves, both travel directions, relevant jaw positions, and any gantry or dose-rate condition associated with the original error. Record results numerically so that future drift can be identified.

The return-to-service decision should be documented and authorised under local policy. Depending on the fault, this may require MLC positional accuracy, picket-fence, output, imaging, end-to-end, or patient-specific verification. The scope should reflect the risk, not simply the convenience of the quickest available test.

Before releasing the unit, review any patients treated since the last known good check. Search for repeated warnings, unusual treatment times, interrupted arcs, failed QA, or manual intervention. If a retrospective review identifies possible clinical impact, escalate promptly through the department’s incident-management process.

A concise service record should include the symptoms, affected patients or plans, diagnostic tests, vendor findings, repairs, calibration data, physics approval, and date of clinical release. Clear records are particularly valuable when a regional Australian service must coordinate between local staff, a remote vendor team, and a visiting engineer.

Build a Repeatable Response

A department is more resilient when MLC troubleshooting is rehearsed rather than improvised. Define who stops treatment, who contacts the vendor, who assesses delivered dose, who informs the radiation oncologist, and who approves return to service. Include after-hours arrangements for sites where the physicist or engineer may not be immediately available.

Useful safeguards include:

• Trend leaf-position and picket-fence results instead of reviewing them only as pass or fail
• Maintain a current inventory of MLC-related interlocks, service contacts, and approved tests
• Keep validated backup treatment approaches for urgent patients
• Review complex VMAT and stereotactic plans with risk-based acceptance criteria
• Train therapists and physicists to preserve logs before resetting the system

The equipment market in Australia includes a mix of older and newer linacs, refurbished systems, and different vendor platforms. A procedure written for one controller may not transfer safely to another. Local work instructions should identify model-specific limitations, software versions, calibration dependencies, and the availability of replacement parts.

Use incident reviews to improve detection rather than assign blame. A near miss involving a leaf-position warning may reveal a documentation gap, unclear handover, inadequate trend analysis, or excessive dependence on a single machine. Share lessons across the service, including with staff in Brisbane, Perth, Adelaide, and smaller regional centres where staffing and equipment redundancy differ.

A malfunctioning MLC should be treated as a controlled clinical event: stop, preserve evidence, classify the failure, test systematically, quantify any dose impact, and release the machine only when the evidence supports safe operation. Apply this workflow to the next warning or failed QA result, and document each decision in a way that supports patients, staff, and the long-term reliability of the radiotherapy service.