|
The customer reported that the precise treatment table was moving after the release of the assisted set-up (asu) button on the function keypad (fkp).On site physics determined that the table was stationary but the readout was changing by greater than 1cm.There is no way to positively determine if any patients were mistreated however it is possible that a mistreatment occurred.The concern is that if only the position readout is moving then the patient could have been incorrectly repositioned to compensate for a readout error.The site has had a new lateral drive assembly, new ltu and lcu pcbs, and ai12/scc in the icca replaced the table has been calibrated and no further issues have been reported.
|
|
This case relates to elekta precise treatment table s/n (b)(4) attached to elekta linac s/n (b)(4) (approximate date of manufacture: november 2012).The manufacturer's final investigation concluded that the likelihood of a mistreatment occurring was very low.No field safety corrective action was required; the issue was determined to be a hardware failure restricted to the reporting customer site, and as a consequence does not impact the installed base, current or future shipments.*investigation details on (b)(6) 2015 - the therapist reported that the table was moving after releasing asu on fkp - using the ratm control via xvi and getting red out of tolerance markers on the lateral axis only.Elekta fse was contacted and explained that it had to be determined if the table was actually moving or if the display was changing while the table was stationary.Physics determined the table was stationary but that the readout was changing by greater than 1cm.Customer was advised to stop treatments.Onsite engineers swapped the ai12-b card in the icca to a different location and checked grub screws on the lateral pots.After completion physics performed table shift qa using a pentaguide phantom with no issues found.Elekta ordered replacement scc icca-a, scc icca-b and an ai12 pcb.On (b)(6) 2015 - installed scc and ai12 cards in icca.Physics again ran 2x table qa with no issues found, so treatments were resumed and no issues reported for the remainder of the day.On (b)(6) 2015 - therapist reported to onsite engineers that the table was hunting laterally during morning qa.Onsite engineers observed treatments and after 4 patients the readout began changing after the asu buttons were released.Elekta advised the customer to stop treatments until further troubleshooting was complete.Onsite engineers inspected and re-connected all cable connections from the lateral drive to the interface cabinet.No issues were noted.Physics ran qa and decided to attempt clinical treatment.The first patient was imaged and shift obtained and again the lateral readout was changing after asu was released.Hospital physicists were strongly advised to stop treatments.Physics ran 10x qa checks without any issues, but were not satisfied that this was a permanent fix so requested replacing the whole lateral drive mechanism.Elekta fse ordered the lateral drive assembly, ltu and lcu pcb's.On (b)(6) 2015 - installed the new lateral drive assembly, ltu, and lcu pcb's.Learned lateral pots and calibrated lateral axis.Physics ran qa and approved clinical treatments.On (b)(6) 2015 - no further issues were reported.A telephone conference with elekta engineers confirmed that the drives and pots removed from the lateral assembly were all correctly secured.The collective opinion of the engineers was that the problem was being mis-reported by the therapists, and that there was no [?]hunting' during the asu, only a readout drift after the asu was released.The most likely root cause is an instability in either the fnaypss or fnbypss analog signal.The suspect faulty components were returned from site to elekta, for the purpose of investigation into the root cause.The effect of the issue upon a workflow using mosaiq's couch move assistant (cma) was investigated.It was determined that the effect would be the same as that for xvi.The log files of the reporting site was examined and they indicate that xvi is moving the table following a scan.Three weeks after the last part was replaced on site, the problem resurfaced - the onsite engineer replaced pcb 835 and cable 45.There have not been any reports of drifting since (b)(6) 2015.Replacement of pcb 835 and cable 45 suggested that the origin was a capacitive/resistive effect caused by cable or connector exhibiting a momentary open circuit or high impedance circuit.An elekta internal meeting held on (b)(6) 2015 concluded that an intermittent high impedance in the fine-pot readback loop is the most likely cause of the observed error; and that pss movement is likely to be a contributory factor.The meeting also concluded that the fault fails safe in that it results in a false negative and never a false positive: high impedance will cause fine-pot reading to drift towards zero with no physical movement.The meeting also highlighted that no reported mv or xvi terminations were logged.This supported the theory that the fault becomes manifest during pss movement and not as a result of releasing the asu controls.The meeting created one action: to simulate the fault in a test shelter and use the result to corroborate the root cause assumptions.In addition all suspect faulty components have been requested to be returned to elekta ltd.(crawley) for examination.On wednesday (b)(6) 2015 an experiment was conducted in an elekta test shelter to corroborate the assumptions made at the previous day's meeting.The outcome of the experiment in short corroborated the assumptions made regarding the root cause being an intermittent high impedance in the fine-pot readback loop.Potential consequences there is no way to positively determine if any patients were mistreated but it is possible that a mistreatment occurred.The concern is that if only the position readout is moving then the patient could have been incorrectly repositioned to compensate for a readout error.An analysis of the intellimax rmx log termination data did not show any evidence of treatment terminations due to table positional errors.The conclusion was that no treatments were affected during the beam delivery phase.The initiating event is most likely a hardware failure between the ai board and the table lateral potentiometer(s).The root cause was assumed to be an intermittent high impedance in the fine-pot feedback loop.This has been corroborated by experimental data.The sequence of events is summarised as: pre-treatment setup asu is performed to correctly position the patient for treatment.Asu controls are released and lateral shift is observed on either xvi, mosaiq and/or tcs.In the case of xvi or mosaiq an out of tolerance lateral position is observed and interpreted as an actual pss move.Another asu is performed moving the table away from the correct position for treatment.Lateral position is stable and new (incorrect) position is accepted for treatment.Patient is potentially mistreated.Note: if however the drift is not noticed by the user in step c), this is a fail-safe scenario, and user continues with workflow as normal.Severity of the hazard/accident the magnitude of the drift in the readout errors is reported as 'greater than 1cm' which is below the error detection limit of the control system (given as 14.8mm).Depending on the treatment site and dose delivered, the severity would have the potential to be major in the case of stereotactic treatment techniques or with high dose close to critical structures.Risk is the product of the severity and likelihood, its current tolerability/acceptability.The two workflow scenarios considered gave rise to the following levels of estimated risk: workflow 1: asu inside of treatment room has an estimated risk of major x incredible = rv3.The risk has been estimated as being comparable with other risks that are already accepted.Workflow 2: asu outside of treatment room has an estimated risk of major x incredible = rv3.The risk has been estimated as being comparable with other risks that are already accepted.Compliance the system is designed to meet the requirements of iec60601-2-1 during normal operation, fault conditions and foreseeable misuse.There is no evidence to suggest that the current design does not continue to meet the requirements of iec60601-2-1.The interlocks that would normally satisfy essential requirements as a result of unplanned movement are unlikely to activate before unacceptable tolerances are reached due to the nature of the set-up and not as a result of a fault.There is every reason to believe that if an actual unplanned movement has occurred then termination would have resulted.Movement can only be initiated before radiation or after termination by the use of two switches, one of which is hard-wired.Education in terms of training and user instruction have already been highlighted final checks to be performed as part of any planned movements.The risk analysis has confirmed that the user would have understood the clinical significant of couch displacement, displayed by the tolerance markers and not to simply overridden this safe limit.It is not considered that this elekta case raises any compliance issues.Conclusion although the definitive root cause could not be established, the issue was likely to have arisen from an intermittent high impedance in the fine-pot feedback loop.This hardware failure was considered to be low risk and was restricted to the reporting customer site, and as a consequence does not impact the installed base, current or future shipments.
|
