Diagnosing atmospheric communication of a sealed monitor chamber

Abstract Daily output variations of up to ±2% were observed for a protracted time on a Varian TrueBeam® STx; these output variations were hypothesized to be the result of atmospheric communication of the sealed monitor chamber. Daily changes in output relative to baseline, measured with an ionization chamber array (DQA3) and the amorphous silicon flat panel detector (IDU) on the TrueBeam®, were compared with daily temperature‐pressure corrections (P TP) determined from sensors within the DQA3. Output measurements were performed using a Farmer® ionization chamber over a 5‐hour period, during which there was controlled variation in the monitor chamber temperature. The root mean square difference between percentage output change from baseline measured with the DQA3 and IDU was 0.50% over all measurements. Over a 7‐month retrospective review of daily changes in output and P TP, weak correlation (R 2 = 0.30) was observed between output and P TP for the first 5 months; for the final 2 months, daily output changes were linearly correlated with changes in P TP, with a slope of 0.84 (R 2 = 0.89). Ionization measurements corrected for ambient temperature and pressure during controlled heating and cooling of the monitor chamber differed from expected values for a sealed monitor chamber by up to 4.6%, but were consistent with expectation for an air‐communicating monitor chamber within uncertainty (1.3%, k = 2). Following replacement of the depressurized monitor chamber, there has been no correlation between daily percentage change in output and P TP (R 2 = 0.09). The utility of control charts is demonstrated for earlier identification of changes in the sensitivity of a sealed monitor chamber.


| INTRODUCTION
Medical linear accelerators use ionization chambers positioned within the beam path to monitor the radiation fluence produced and terminate an irradiation when the programmed fluence has been delivered. The fluence measured by the monitor chambers is related to the dose delivered to a specified point for a reference set of irradiation conditions as defined by a clinical reference dosimetry protocol between December 20, 2017 and February 23, 2018, the standard deviation in daily measurements of dose per monitor unit relative to baseline increased to 0.84%. This increased standard deviation in daily measurements of dose per monitor unit was determined to be the result of atmospheric communication of the sealed monitor chamber. While long-term output trend analyses for linear accelerators with sealed monitor chambers have previously been reported, 3,4 atmospheric communication of a sealed monitor chamber has only been reported for an obsolete linear accelerator model with limited guidance on the process for diagnosis of atmospheric communication. 5 The purpose of this work was to develop a process for the diagnosis of atmospheric communication of a sealed monitor chamber.

| MATERIALS AND METHODS
To verify that the increased variation in daily measurements of dose per monitor unit with the DQA3 was not due to measurement error, independent measurements of the change in dose per monitor unit relative to baseline were made daily with the amorphous silicon flat panel detector (IDU) on the TrueBeam® STx using the beam constancy check of the TrueBeam® Machine Performance Check (MPC) application beginning on December 11, 2017. Consistency of MPC output constancy measurements with ionization chamber measurements has previously been demonstrated. 6 The DQA3 has built-in thermistors and pressure sensors to correct measurements for atmospheric conditions. A visual review of trended daily temperature-pressure corrections (P TP ) as determined by the DQA3 revealed that, by December 20, 2017, the daily variations in dose per monitor unit were consistent in magnitude with the variations in P TP , suggesting that the monitor chamber was communicating with the atmosphere.
To test for atmospheric communication of the monitor chamber, ionization measurements were performed using a PTW 23333 Farmer® ionization chamber (PTW, Freiburg, Germany) during controlled temperature variation in the monitor chamber. The Farmer® ionization chamber was positioned in a Solid Water® (Sun Nuclear Corporation, Melbourne, FL) phantom at a depth of 10 cm with 5 cm of backscatter. Ionization measurements for irradiations of 100 MU were completed using a PTW Unidos E electrometer over a 5-h period during which the temperature of the monitor chamber was initially increased using a heat gun, then decreased using air conditioning and a fan. The temperatures of both the monitor chamber and the Solid Water® phantom were measured prior to each irradiation using a noncontact infrared thermometer. Each temperature measurement was repeated three times in succession across the visible surfaces of the monitor chamber and the Solid Water® phantom to provide an estimate of uncertainty due to temperature gradients and repeatability of the thermometer. All measurements were performed using the 6 MV beam energy; however, the increased variations in daily output measurements that motivated this investigation were observed for all energies.
Changes in the ionization measurements corrected for ambient temperature and pressure were compared with expectation assuming (a) the monitor chamber was sealed and (b) the monitor chamber was air communicating. For the assumption of a sealed monitor chamber, the corrected ionization measurements should be constant.
For an air-communicating monitor chamber, the corrected ionization measurements should be directly proportional to changes in P TP since a decrease in ambient air density (i.e., increase in P TP ) requires a greater fluence at the monitor chamber to produce a given quantity of ionization within the monitor chamber (i.e., monitor unit).
Based on the results of these measurements, the monitor chamber

| RESULTS AND DISCUSSION
The root mean square difference between percentage change in dose per monitor unit measured with the DQA3 and the IDU is 0.50% over all measurements prior to the replacement of the monitor chamber. This agreement between DQA3 and IDU measurements provides verification of the observed increase in the standard deviation of DQA3 measurements (from 0.36% to 0.84%) that motivated this investigation. Figure 2 shows the percentage change from baseline in the dose per monitor unit measured daily using the  has been no correlation between trend-corrected DQA3 measurements and atmospheric changes (Fig. 6), and the standard deviation in trend-corrected DQA3 measurements has been reduced to 0.18%. F I G . 6. Percentage change in output from baseline measured daily using an ionization chamber array plotted versus the percentage change from the temperature-pressure correction (P TP ) at the time of baseline. The monitor chamber was replaced on 2/23/2018. F I G . 7. Average (a) and range (b) charts for the daily output measurements with the ionization chamber array (DQA3) before replacement of the monitor chamber using a subgroup size of one. Consecutive results in the average chart were used to calculate the range. Control limits indicated by the thin solid lines were calculated using the first 10 data points. The yellow and red lines indicate institutional limits for scheduled and immediate action, respectively.
F I G . 8. Average (a) and range (b) charts for the daily output measurements with the ionization chamber array (DQA3) after replacement of the monitor chamber using a subgroup size of one. Consecutive results in the average chart were used to calculate the range. Effects of output calibration events and the increasing trend in output have been removed. Control limits indicated by the thin solid lines were calculated using the first ten data points. The yellow and red lines indicate institutional limits for scheduled and immediate action, respectively