# Toward Noninvasive High‐Resolution In Vivo pH Mapping in Brain Tumors by 31P‐Informed deepCEST MRI

## TL;DR

This study introduces a deep learning method to create high-resolution brain tumor pH maps using MRI data, improving on traditional techniques with faster scans and better detail.

## Contribution

A neural network is proposed to predict high-resolution 31P-pHi maps from CEST data, enabling noninvasive pH imaging with improved spatial resolution and scan time.

## Key findings

- The predicted deepCEST pHi maps showed general correspondence with measured 31P-pHi values, with an RMSE of 0.04 pH units in tumor regions.
- High-resolution predictions revealed tumor heterogeneity and features not visible in conventional CEST data.
- The model captures unique pH information and is not simply a T1 segmentation, offering a promising approach for 3D pH imaging.

## Abstract

The intracellular pH (pHi) is critical for understanding various pathologies, including brain tumors. While conventional pHi measurement through 31P‐MRS suffers from low spatial resolution and long scan times, 1H‐based APT‐CEST imaging offers higher resolution with shorter scan times. This study aims to directly predict 31P‐pHi maps from CEST data by using a fully connected neuronal network. Fifteen tumor patients were scanned on a 3‐T Siemens PRISMA scanner and received 1H‐based CEST and T1 measurement, as well as 31P‐MRS. A neural network was trained voxel‐wise on CEST and T1 data to predict 31P‐pHi values, using data from 11 patients for training and 4 for testing. The predicted pHi maps were additionally down‐sampled to the original the 31P‐pHi resolution, to be able to calculate the RMSE and analyze the correlation, while higher resolved predictions were compared with conventional CEST metrics. The results demonstrated a general correspondence between the predicted deepCEST pHi maps and the measured 31P‐pHi in test patients. However, slight discrepancies were also observed, with a RMSE of 0.04 pH units in tumor regions. High‐resolution predictions revealed tumor heterogeneity and features not visible in conventional CEST data, suggesting the model captures unique pH information and is not simply a T1 segmentation. The deepCEST pHi neural network enables the APT‐CEST hidden pH‐sensitivity and offers pHi maps with higher spatial resolution in shorter scan time compared with 31P‐MRS. Although this approach is constrained by the limitations of the acquired data, it can be extended with additional CEST features for future studies, thereby offering a promising approach for 3D pH imaging in a clinical environment.

31P‐MRS is a noninvasive method for measuring intracellular pH. This study presents a 31P‐informed deep CEST approach for generating high resolved pHi data by using proton‐based APT‐CEST input data. The higher resolution enables contrast features, which are not a pure T1 segmentation and differ from conventional CEST metrics