Integrated PET/CT in the preoperative staging of lung cancer: A prospective comparison of CT, PET and integrated PET/CT

A machine learning tool to improve prediction of mediastinal lymph node metastases in non-small cell lung cancer using routinely obtainable [18F]FDG-PET/CT parameters

BACKGROUND

In patients with non-small cell lung cancer (NSCLC), accuracy of [18F]FDG-PET/CT for pretherapeutic lymph node (LN) staging is limited by false positive findings. Our aim was to evaluate machine learning with routinely obtainable variables to improve accuracy over standard visual image assessment.

METHODS

Monocentric retrospective analysis of pretherapeutic [18F]FDG-PET/CT in 491 consecutive patients with NSCLC using an analog PET/CT scanner (training + test cohort, n = 385) or digital scanner (validation, n = 106). Forty clinical variables, tumor characteristics, and image variables (e.g., primary tumor and LN SUVmax and size) were collected. Different combinations of machine learning methods for feature selection and classification of N0/1 vs. N2/3 disease were compared. Ten-fold nested cross-validation was used to derive the mean area under the ROC curve of the ten test folds ("test AUC") and AUC in the validation cohort. Reference standard was the final N stage from interdisciplinary consensus (histological results for N2/3 LNs in 96%).

RESULTS

N2/3 disease was present in 190 patients (39%; training + test, 37%; validation, 46%; p = 0.09). A gradient boosting classifier (GBM) with 10 features was selected as the final model based on test AUC of 0.91 (95% confidence interval, 0.87-0.94). Validation AUC was 0.94 (0.89-0.98). At a target sensitivity of approx. 90%, test/validation accuracy of the GBM was 0.78/0.87. This was significantly higher than the accuracy based on "mediastinal LN uptake > mediastinum" (0.7/0.75; each p < 0.05) or combined PET/CT criteria (PET positive and/or LN short axis diameter > 10 mm; 0.68/0.75; each p < 0.001). Harmonization of PET images between the two scanners affected SUVmax and visual assessment of the LNs but did not diminish the AUC of the GBM.

CONCLUSIONS

A machine learning model based on routinely available variables from [18F]FDG-PET/CT improved accuracy in mediastinal LN staging compared to established visual assessment criteria. A web application implementing this model was made available.

The added value of18F-FDG PET/CT in staging non-small cell lung cancer

Background

Non-small cell lung cancer (NSCLC) accounts for about 80% of all lung cancers. The current criteria for its staging are based on the TNM system that determines treatment options and predicts survival rate in patients. The aim of the study was to evaluate the diagnostic accuracy of 18F-fluorodeoxyglucose (18F-FDG) positron emission tomography PET/CT in staging of NSCLC patients.

Methods

A retrospective study was conducted. We reviewed the CT and PET/CT examinations done in our institution on pathologically proven patients of NSCLC, in the period between October 2018 and end of July 2019.

Results

A total of 40 patients were evaluated with the age ranging from 37 to 77 years old, and the mean was 55.63 years (SD ± 10.29). There were 31 male cases and 9 female cases. When we compared contrast enhanced CT (CECT) to PET-CT for staging, PET-CT helped upstage disease in 10 of 40 patients (25%) and downstage in 3 of 40 patients (7.5%).

Conclusion

PET/CT is a useful imaging tool in initial staging of the newly diagnosed patients with NSCLC. It is better thаn СT alone fоr detection of malignant lesions for accurate staging. It can change the strategy of treatment according to its findings.

A Novel Adaptive PET/CT Image Fusion Algorithm

Background:

Lung cancer is one of the common malignant tumors. The successful diagnosis of lung cancer depends on the accuracy of the image obtained from medical imaging modalities.

Objective:

The fusion of CT and PET is combining the complimentary and redundant information both images and can increase the ease of perception. Since the existing fusion method sare not perfect enough, and the fusion effect remains to be improved, the paper proposes a novel method called adaptive PET/CT fusion for lung cancer in Piella framework.

Methods:

This algorithm firstly adopted the DTCWT to decompose the PET and CT images into different components, respectively. In accordance with the characteristics of low-frequency and high-frequency components and the features of PET and CT image, 5 membership functions are used as a combination method so as to determine the fusion weight for low-frequency components. In order to fuse different high-frequency components, we select the energy difference of decomposition coefficients as the match measure, and the local energy as the activity measure; in addition, the decision factor is also determined for the high-frequency components.

Results:

The proposed method is compared with some of the pixel-level spatial domain image fusion algorithms. The experimental results show that our proposed algorithm is feasible and effective.

Conclusion:

Our proposed algorithm can better retain and protrude the lesions edge information and the texture information of lesions in the image fusion.

PET-CT for assessing mediastinal lymph node involvement in patients with suspected resectable non-small cell lung cancer

BACKGROUND

A major determinant of treatment offered to patients with non-small cell lung cancer (NSCLC) is their intrathoracic (mediastinal) nodal status. If the disease has not spread to the ipsilateral mediastinal nodes, subcarinal (N2) nodes, or both, and the patient is otherwise considered fit for surgery, resection is often the treatment of choice. Planning the optimal treatment is therefore critically dependent on accurate staging of the disease. PET-CT (positron emission tomography-computed tomography) is a non-invasive staging method of the mediastinum, which is increasingly available and used by lung cancer multidisciplinary teams. Although the non-invasive nature of PET-CT constitutes one of its major advantages, PET-CT may be suboptimal in detecting malignancy in normal-sized lymph nodes and in ruling out malignancy in patients with coexisting inflammatory or infectious diseases.

OBJECTIVES

To determine the diagnostic accuracy of integrated PET-CT for mediastinal staging of patients with suspected or confirmed NSCLC that is potentially suitable for treatment with curative intent.

SEARCH METHODS

We searched the following databases up to 30 April 2013: The Cochrane Library, MEDLINE via OvidSP (from 1946), Embase via OvidSP (from 1974), PreMEDLINE via OvidSP, OpenGrey, ProQuest Dissertations & Theses, and the trials register www.clinicaltrials.gov. There were no language or publication status restrictions on the search. We also contacted researchers in the field, checked reference lists, and conducted citation searches (with an end-date of 9 July 2013) of relevant studies.

SELECTION CRITERIA

Prospective or retrospective cross-sectional studies that assessed the diagnostic accuracy of integrated PET-CT for diagnosing N2 disease in patients with suspected resectable NSCLC. The studies must have used pathology as the reference standard and reported participants as the unit of analysis.

DATA COLLECTION AND ANALYSIS

Two authors independently extracted data pertaining to the study characteristics and the number of true and false positives and true and false negatives for the index test, and they independently assessed the quality of the included studies using QUADAS-2. We calculated sensitivity and specificity with 95% confidence intervals (CI) for each study and performed two main analyses based on the criteria for test positivity employed: Activity > background or SUVmax ≥ 2.5 (SUVmax = maximum standardised uptake value), where we fitted a summary receiver operating characteristic (ROC) curve using a hierarchical summary ROC (HSROC) model for each subset of studies. We identified the average operating point on the SROC curve and computed the average sensitivities and specificities. We checked for heterogeneity and examined the robustness of the meta-analyses through sensitivity analyses.

MAIN RESULTS

We included 45 studies, and based on the criteria for PET-CT positivity, we categorised the included studies into three groups: Activity > background (18 studies, N = 2823, prevalence of N2 and N3 nodes = 679/2328), SUVmax ≥ 2.5 (12 studies, N = 1656, prevalence of N2 and N3 nodes = 465/1656), and Other/mixed (15 studies, N = 1616, prevalence of N2 to N3 nodes = 400/1616). None of the studies reported (any) adverse events. Under-reporting generally hampered the quality assessment of the studies, and in 30/45 studies, the applicability of the study populations was of high or unclear concern.The summary sensitivity and specificity estimates for the 'Activity > background PET-CT positivity criterion were 77.4% (95% CI 65.3 to 86.1) and 90.1% (95% CI 85.3 to 93.5), respectively, but the accuracy estimates of these studies in ROC space showed a wide prediction region. This indicated high between-study heterogeneity and a relatively large 95% confidence region around the summary value of sensitivity and specificity, denoting a lack of precision. Sensitivity analyses suggested that the overall estimate of sensitivity was especially susceptible to selection bias; reference standard bias; clear definition of test positivity; and to a lesser extent, index test bias and commercial funding bias, with lower combined estimates of sensitivity observed for all the low 'Risk of bias' studies compared with the full analysis.The summary sensitivity and specificity estimates for the SUVmax ≥ 2.5 PET-CT positivity criterion were 81.3% (95% CI 70.2 to 88.9) and 79.4% (95% CI 70 to 86.5), respectively.In this group, the accuracy estimates of these studies in ROC space also showed a very wide prediction region. This indicated very high between-study heterogeneity, and there was a relatively large 95% confidence region around the summary value of sensitivity and specificity, denoting a clear lack of precision. Sensitivity analyses suggested that both overall accuracy estimates were marginally sensitive to flow and timing bias and commercial funding bias, which both lead to slightly lower estimates of sensitivity and specificity.Heterogeneity analyses showed that the accuracy estimates were significantly influenced by country of study origin, percentage of participants with adenocarcinoma, (¹⁸F)-2-fluoro-deoxy-D-glucose (FDG) dose, type of PET-CT scanner, and study size, but not by study design, consecutive recruitment, attenuation correction, year of publication, or tuberculosis incidence rate per 100,000 population.

AUTHORS' CONCLUSIONS

This review has shown that accuracy of PET-CT is insufficient to allow management based on PET-CT alone. The findings therefore support National Institute for Health and Care (formally 'clinical') Excellence (NICE) guidance on this topic, where PET-CT is used to guide clinicians in the next step: either a biopsy or where negative and nodes are small, directly to surgery. The apparent difference between the two main makes of PET-CT scanner is important and may influence the treatment decision in some circumstances. The differences in PET-CT accuracy estimates between scanner makes, NSCLC subtypes, FDG dose, and country of study origin, along with the general variability of results, suggest that all large centres should actively monitor their accuracy. This is so that they can make reliable decisions based on their own results and identify the populations in which PET-CT is of most use or potentially little value.

Meta-analysis of the additional value of integrated 18FDG PET-CT for tumor distant metastasis staging: comparison with 18FDG PET alone and CT alone

BACKGROUND

We performed a meta-analysis to compare the diagnostic performance of integrated (18)FDG PET-CT with that of (18)FDG PET alone and CT alone for the overall assessment of tumor distant metastases.

METHODS

We performed a meta-analysis of 11 available articles (1107 patients). We calculated sensitivities, specificities, positive likelihood ratios, and negative likelihood ratios for integrated PET-CT, PET alone, and CT alone, respectively. We also constructed summary receiver operating characteristic curves for integrated PET-CT, PET alone, and CT alone, respectively.

RESULTS

Across 10 comparative studies (1058 patients), integrated PET-CT has higher sensitivity (0.95 vs 0.85) and similar specificity (0.96 vs 0.95) with PET alone. Across 7 comparative studies (745 patients), integrated PET-CT has higher sensitivity (0.97 vs 0.80) and similar specificity (0.97 vs 0.94) with CT alone.

CONCLUSION

Integrated PET-CT is the most sensitive and accurate modality for tumor distant metastasis staging. Integrating PET with CT may tend to improve the sensitivity than PET alone and CT alone.