Quality control of superficial hyperthermia by treatment evaluation

Figures & data

Figure 1. (A) Input–output schematic diagram of a superficial hyperthermia target volume. The patient-related factors are represented by the vertical arrow. The outputs marked * are subjective and not recorded by the data acquisition system. (B) Closed-loop control in superficial hyperthermia where the human controller interprets tissue temperatures and patient feedback and steers the water bolus temperature and applicator powers. Numbers in italic font represent the signal width and are specific to the circumstances at the Erasmus MC.

Figure 2. A 2 × 3 LCA array and water bolus (WB) placed on the chest of a patient.

Figure 3. Screenshot of the graphical user interface for power steering. The measured forward power is indicated in large font and the power set point in small font. The reflected powers are represented graphically in a colour code (green <5%, yellow 5–10%, orange 10–15%, red >15% reflected power).

Figure 4. Screenshot of the temperature monitoring graphical user interface. (A) Temperature–time graphs for the sensors below each applicator are shown. (B) The measured temperatures are mapped onto a sketch that includes anatomical features, the location of the tumour and scar tissue, apertures of the applicators, catheter tracks and margins of the radiation field. For each measurement point the tissue type, depth, temperature and temperature difference over 20 s is also shown.

Figure 5. Example treatment evaluation sheet. Ch, channel; ss, steady-state; int., interstitial; sup., superficial; NaN, not-a-number (no data available).

Table I.  Power steering statistics based on a per applicator analysis of all treatments conducted in 2009 at Erasmus MC (196 treatments, 920 applicator·treatments).

Steering parameter	Steady-state mean power (W)	Number of switch off events	Time off (min)	Number of steering actions
Mean	56.4	3.7	2.7	17.8
10th percentile	30	1	0.1	9
25th percentile	42	1	0.4	12
Median	55	3	1.4	16
75th percentile	69	5	3.0	21
90th percentile	85	7	5.2	29

Figure 6. Characteristic shapes of the power trend line: (A) normal, (B) peak at beginning of treatment, (C) continuous increase and (D) decrease at the end of treatment.

Figure 7. Characteristic shapes of the water bolus trend line: (A) stable and normal, (B) low temperature at the start, (C) wavy and (D) steering.

Figure 8. Characteristic shapes of the interstitial temperature trend line: (A) normal, stable, (B) increasing towards the end, (C) decline towards the end, (D) maximum in the early stages and (E) dip in the steady-state period.

Figure 9. Examples of characteristic temperature–depth plots: (A) all temperatures therapeutic, (B) mix of therapeutic and sub-therapeutic temperatures, (C) sub-therapeutic temperatures, (D) available data only in the top layer with poor coverage at depth, (E) no interstitial data available and (F) relatively high surface and low interstitial temperatures indicating a too high water bolus temperature.

Table II.  Characteristics of typical power trend line shapes, the related checks and possible actions for the next treatment.

Trend line shape	Characteristics	Checks and possible actions
Normal	The trend line rises to above the average within the first 20 min and remains there	–
Peak at beginning of treatment	Peak in the trend line within the first half of the treatment	Check if a too high start value or relatively large increments in power triggered early limitation events
Continuous increase	The total power rises until the end of treatment	Possibly the initial power and/or power increments were too modest. It may also reflect the willingness to treat in the upper therapeutic range with a tolerant patient
Decrease at the end of treatment	The trend line slopes downwards in the second stage of the treatment	This implies frequent off-switching due to pain complaints or limiting temperatures. The power build up in the first half may have been too fast

Table III.  Characteristics of typical water bolus temperature trend line shapes, the related checks and possible actions for the next treatment.

Trend line shape	Characteristics	Checks and possible actions
Stable and normal	The trend line shows little deviation from the mean	–
Low temperature at the start	The temperature is far below average in the first 10–20 min	Probably the water bolus was not heated to the target temperature before the treatment, or the temperature probe was not correctly placed in the inlet
Wavy	The trend line fluctuates up and down	Possibly the treatment was interrupted and the water bolus temporarily removed from the patient. Otherwise, check for malfunctioning of the equipment
Steering	The trend line reflects the steering actions	Check against the temperature data, if steering was effective

Table IV.  Characteristics of typical interstitial temperature trend line shapes, the related checks and possible actions during treatment evaluation.

Trend line shape	Characteristics	Checks and possible actions
Normal, stable	After 15–20 min the trend line settles to a steady-state at a temperature between 40°–43°C	–
Increasing towards the end	The trend line has a positive slope throughout the treatment	Consider a higher start power or faster power build up
Decline towards the end	The average interstitial temperature drops in the last 20 min	Possibly limiting factors are handled with difficulty towards the end of treatment
Maximum in the early stages	An initially fast rise is followed by a steady decline to or below the average temperature	Consider a more gentle start of the treatment such that the temperature will not drop due to too early pain complaints
Dip in the steady-state period	Dips in the holding state	Dips often reflect severe limitations such as power off for several min, which may take more than 10 min to compensate for

Figure 10. Schematic representation of the four main levels of controllability and observability in parts of the hyperthermia target volume (HTV). The blocks marked ∫ represent the tissue dynamics of a HTV subsection. From top to bottom: controllable and observable, controllable and poorly observable, poorly controllable and observable, poorly controllable and poorly observable.