Low-dose ultrahigh-resolution PCCT enhances subsolid nodule characterization

Abstract

Purpose: To characterize invasion-associated CT features in pulmonary subsolid nodules using low-dose ultrahigh-resolution (UHR) photon-counting CT (PCCT) images and evaluate UHR's diagnostic superiority over standard high-resolution (HR) images.

Methods: Patients with subsolid lung adenocarcinoma were recruited for chest scan on PCCT to obtain UHR and standard HR images between November 2023 and May 2024. Nodule characteristics were visually assessed and histogram features were extracted from each nodule. Image quality and radiation dose at previous energy-integrating detector CT (EID-CT) of 30 patients were compared with those of PCCT. Differences between UHR and standard HR, PCCT and EID-CT were compared using paired McNemar-test or paired Wilcox-test.

Results: One hundred and eighty-four patients with 203 subsolid nodules were collected including 77 precursors, 77 minimally invasive adenocarcinoma (MIA) and 49 IA. UHR significantly outperformed standard HR in revealing CT findings including larger nodular diameter and solid-component diameter, more frequency of heterogeneous attenuation, lobulation, bubble-like sign, air bronchogram, pleural indentation and vascular sign (all P < 0.05). Additionally, UHR images exhibited significantly greater value in histogram-derived parameters compared to standard HR images (all P < 0.05), except for "Median," "Minimum." Furthermore, the radiation dose in PCCT was half of that in EID-CT (effective dose: 1.32 ± 0.27 vs. 3.85 ± 1.65/mSv, P < 0.001. CDTI_vol: 2.97 ± 0.53 vs. 6.90 ± 2.97/mGy, P < 0.001), with image quality significantly better in PCCT.

Conclusion: The UHR protocol on PCCT provides a magnified perspective to reveal CT characteristics of invasive growth in subsolid LUAD, previously undetectable on standard HR images, achieving halved radiation dose and better image quality than EID-CT.

Keywords: Lung adenocarcinoma; Photon-counting CT; Subsolid nodules; Ultrahigh resolution.

Fig. 1

The workflow of patients’ enrollment

Fig. 2

An example of image quality assessment. Signal-to-noise ratio (SNR) = ∣CT _lung = (-921HU) / SD _lung = (39HU) ∣ = 23.6, Contrast-to-noise ratio (CNR) = ∣ (CT _artery = (42HU) – CT _muscle = (52HU)) / SD _fat = (50HU) ∣ = 0.2

Fig. 3

Paired boxplots comparing nodular diameter, solid-component diameter, CT value, and its standard deviation (SD) between ultrahigh-resolution (UHR) and standard HR images across all LUAD subtypes (A), in the pre-IA subtype (B) and in the IA subtype (C)

Fig. 4

Bar charts comparing radiographic signs between UHR and standard HR images across all LUAD subtypes (A), in the pre-IA subtype (B) and in the IA subtype (C), respectively

Fig. 5

A An elderly female pathologically confirmed with IA in the right lower lobe. UHR images revealed a subtle solid component on both lung and mediastinal window settings (triangle), however, this critical sign was obscured on standard HR images, leading to misclassification of a pure ground-glass nodule (pGGN) and underestimation of the invasiveness of the LUAD. B A middle-aged woman with AIS in the right middle lobe. On standard HR images, the nodule exhibited a heterogeneous appearance with central subtle solid component (triangle). Contrarily, UHR images revealed this central component corresponds to a microvessel (arrow). An another microvessel was found at nodule periphery (arrow). C an elderly woman with AIS in the right lower lobe. On standard HR images, the peripheral solid microfoci (triangle) was initially classified as a mixed GGN category. However, UHR images revealed this structure corresponds to a microvascular component

Fig. 6

A An elderly male with MIA. On UHR images, the air bronchogram (arrow) is clearly shown within the nodule. In contrast, standard HR images only demonstrate an ill-defined hypoattenuating area adjacent to the nodule. Besides, UHR images enable more precise delineation of the nodular margin (arrowhead) compared to standard HR images. B A middle-aged female with MIA. UHR images more clearly demonstrate lobulation (arrowhead), heterogeneous attenuation, a coarse tumor–lung interface, and microvessels compared to standard HR image. Corresponding H&E image ($\times$ 200 times) shows tumor cell nests and widen lobular septa (※). C An elderly female with IA. Compared to standard HR, UHR images demonstrate enhanced visualization of the bubble-like sign (arrow), nodular margin (arrowhead), and air bronchogram. The bubble-like appearance corresponds to the air-containing area (※) in H&E image ($\times$ 200 times)