Evaluating a New Detector for In Vivo Dosimetry

A new detector for in vivo dosimetry should be judged by how reliably it supports clinical decisions, rather than by sensitivity alone. The central question is whether its reading can identify a meaningful treatment-delivery error under routine conditions, while remaining stable across patients, treatment sites, machines and operators.

For Australian medical physicists, this assessment also needs to fit the realities of busy radiotherapy departments in Sydney, Melbourne, Brisbane, Perth and regional centres. A detector that performs well in a controlled laboratory may be less useful if it adds complex setup steps, requires frequent recalibration or produces results that are difficult to interpret during a full treatment list.

Define The Clinical Purpose

In vivo dosimetry measures radiation delivered to, or emerging from, the patient during treatment. Depending on the technology, the detector may be positioned on the patient surface, placed near the treatment field, attached to an immobilisation device or integrated into the treatment system. Its role may be first-fraction verification, ongoing treatment monitoring, incident investigation or independent confirmation of machine output.

Before testing begins, define the error the system must detect. A detector intended to identify incorrect monitor units has different requirements from one designed to reveal patient-positioning errors, beam interruptions or changes in tissue thickness. Establishing action levels in advance prevents the evaluation from becoming a simple comparison of numerical readings.

A useful clinical endpoint could be the percentage difference between measured and expected dose, the rate of correctly identified delivery errors or the time required to review a result. The detector should add assurance without creating unnecessary alerts that staff learn to ignore.

Build A Robust Measurement Protocol

The evaluation should compare the new device with a method already trusted in the department. Depending on the application, this reference may be a calibrated ion chamber, a traceable diode, an optically stimulated luminescence dosimeter, an electronic portal imaging device or a well-characterised treatment-planning calculation. Agreement should be examined across clinically relevant dose levels rather than at a single nominal setting.

Test repeatability by delivering the same field several times, then assess reproducibility across different days, operators and detector placements. Include open fields, shaped fields, intensity-modulated radiotherapy and volumetric-modulated arc therapy where these techniques are part of local practice. Measurements should cover the dose range used for common breast, prostate, head-and-neck and stereotactic treatments.

Energy dependence, dose-rate response, field-size dependence and angular response can expose weaknesses that are hidden during basic calibration. For surface or entrance measurements, investigate the effect of buildup material, oblique incidence, air gaps and patient curvature. Record environmental conditions if the detector is sensitive to temperature, pressure, humidity or storage history.

Analyse Signal Quality And Uncertainty

A clinically useful detector needs a stable signal-to-noise ratio at the lowest dose that matters for the intended application. Evaluate zero readings, dark signal, drift, saturation, response time and the effect of repeated irradiation. A slow recovery time may be acceptable for offline verification but unsuitable for real-time monitoring.

Uncertainty should be reported as a budget, not reduced to a single percentage without explanation. Include calibration, positioning, phantom or patient anatomy, treatment-machine output, detector response, readout electronics and the reference method. Correlation with a reference detector does not prove accuracy if both systems share the same source of error.

Statistical tools should match the clinical question. Bland–Altman analysis can show bias and limits of agreement, while regression can describe proportional errors. Sensitivity and specificity are important when the detector is used as an alerting system. A practical evaluation should also record false positives, false negatives, failed readings and the percentage of cases requiring manual investigation.

Translate Evidence To Australian Practice

Implementation in Australia should account for state and territory radiation-control requirements, local quality-management systems and the expectations of the Australian Radiation Protection and Nuclear Safety Agency. Facilities should document calibration traceability, acceptance testing, ongoing constancy checks and responsibilities for reviewing abnormal readings. Where equipment is supplied as a medical device, procurement and governance teams should also clarify applicable Therapeutic Goods Administration obligations.

The local service environment matters. A metropolitan department may have access to rapid technical support and several treatment platforms, whereas a regional service may need a detector that can be maintained with limited onsite resources. Australian departments commonly coordinate work across multiple campuses, so data export, cybersecurity, software compatibility and remote review can be as important as the detector’s physical response.

Daily workflow should be tested realistically. Staff may be managing early starts, short appointment intervals and patients who travel considerable distances for treatment. A detector that requires a lengthy warm-up or a separate workstation may be bypassed during a busy morning list. Local purchasing should consider consumable costs, import lead times, warranty arrangements and Australian-dollar pricing rather than relying on overseas demonstrations.

Professional exchange also helps reveal practical limitations. The scientific and social programme preserved in the Banff event archive reflects how medical physicists use meetings to discuss emerging detectors, commissioning experiences and quality-assurance practice beyond formal papers.

Turn Results Into A Safe Decision

Acceptance criteria should be agreed before the final data review. For example, the department might specify maximum bias, repeatability, allowable drift, detection capability for predefined delivery errors and a required percentage of usable clinical measurements. Criteria should distinguish between performance suitable for commissioning and performance suitable for routine patient-specific verification.

A staged introduction is safer than immediate deployment across every treatment unit. Begin with a limited pilot involving representative sites and treatment techniques, compare results with the established verification method, and review every alert. This phase can reveal placement errors, communication gaps and unrealistic thresholds before the detector becomes part of the formal treatment pathway.

The complete record should include the detector serial number, calibration date, software version, treatment parameters, placement method, reference dose, environmental observations and reviewer decision. A clear escalation process is essential: an unexpected reading should trigger technical checks, patient- or plan-specific review, and consultation with the radiation oncologist when clinically appropriate.

Practical Recommendations

• Define the clinical error the detector must identify before selecting test fields or statistical methods.
• Use a calibrated reference method and document its traceability, limitations and uncertainty.
• Test linearity, repeatability, reproducibility, angular response, energy dependence and dose-rate response.
• Include realistic patient geometries, treatment techniques and placement conditions in the protocol.
• Set alert thresholds using clinical consequences, rather than relying only on laboratory variation.
• Pilot the detector on a small number of treatment units and review false alerts before wider adoption.
• Maintain a controlled record of calibration, software changes, maintenance, training and clinical decisions.

The archived COMP 2014 conference site provides useful historical context for the professional setting in which medical physicists share evaluation methods, commissioning experience and quality-assurance standards. Although detector technology has advanced, the need for transparent evidence and practical governance remains constant.

A defensible performance study should finish with a clear operational decision: accept the detector for a defined use, restrict it to investigation, require further testing or reject it for clinical deployment. Apply that decision through a documented local procedure, train the staff who will interpret the readings, and review performance after implementation using real Australian treatment data.