Inter-fractional monitoring of [Formula: see text]C ions treatments: results from a clinical trial at the CNAO facility

Abstract

The high dose conformity and healthy tissue sparing achievable in Particle Therapy when using C ions calls for safety factors in treatment planning, to prevent the tumor under-dosage related to the possible occurrence of inter-fractional morphological changes during a treatment. This limitation could be overcome by a range monitor, still missing in clinical routine, capable of providing on-line feedback. The Dose Profiler (DP) is a detector developed within the INnovative Solution for In-beam Dosimetry in hadronthErapy (INSIDE) collaboration for the monitoring of carbon ion treatments at the CNAO facility (Centro Nazionale di Adroterapia Oncologica) exploiting the detection of charged secondary fragments that escape from the patient. The DP capability to detect inter-fractional changes is demonstrated by comparing the obtained fragment emission maps in different fractions of the treatments enrolled in the first ever clinical trial of such a monitoring system, performed at CNAO. The case of a CNAO patient that underwent a significant morphological change is presented in detail, focusing on the implications that can be drawn for the achievable inter-fractional monitoring DP sensitivity in real clinical conditions. The results have been cross-checked against a simulation study.

Figure 1

View of the INSIDE cart with the DP (beige box) and the PET detectors (white boxes above and below the patient bed) installed in the CNAO treatment room n.1, which has a fixed horizontal beam line. The image shows a view of the first patient treated with ${}^{12}$C ions that has been monitored by means of the DP.

Figure 2

Treatment plan delivered to the patient selected for the MC study. In the horizontal plane the two different fields used to treat the patient can be seen. The black contour represents the planning Clinical Target Volume (CTV).

Figure 3

View of CT1 (left) and CT2 (right). The emptying of the nasal cavities induced by the treatment is highlighted with orange boxes.

Figure 4

Left: The p-value distribution obtained performing the ${\chi }^2$ test between the 1D projections relative to the first measured fraction (16 October) and all the other monitored fractions is shown. Right: Reconstructed fragment emission profiles measured in three different fractions. A clear shape change can be observed due to the nasal cavities emptying.

Figure 5

Left: view of CT1 of PZC. Center: view of CT2 of PZC. In both images the region interested by the morphological change, spotted via the re-evaluation CT, is highlighted in orange. The emptying of the region can be clearly seen. Right: CT2 of PZC with superimposed, in cyan, the distribution of the PCAs belonging to the SPBs that have a p-value below 0.02 when testing the consistency between the first monitored fraction and the last one.

Figure 6

Results of a MC simulation. Left: PCA distributions obtained using respectively CT1 and CT2 as input. The shown SPB has been selected among the ones that have a ${\chi }^2$ p-value below 0.02 (for details see text). The results prove that we are observing statistically significant indications of morphological changes. Right: CT2 of PZC with superimposed, in green, the distribution of the PCAs belonging to the SPBs that have a p-value below 0.02 when testing the consistency between CT1 and CT2 (discrepant SPBs).

Figure 7

Left: Superposition of the PCA maps obtained using PZC CT1 as input for the MC simulations to generate two independent samples (shown in yellow and red). The orange region identifies the maps overlapping. Center: Superposition of the PCAs maps from the MC samples generated using respectively PZC CT1 (yellow) and CT2 (red) as input. The longer range of the particles in CT2 is highlighted by the red area that exceeds the orange (overlapping) region. Right: PCAs maps obtained from the data collected respectively in the first (yellow) and the last (red) treamtent fractions.