Prognostic Indicators and Comparative Treatment Outcomes in High-Risk Thyroid Cancer with Laryngotracheal Invasion
Article information
Abstract
Background
Laryngotracheal invasion occurs in a subset of patients with well-differentiated thyroid cancer (WDTC) and is associated with a poor prognosis. We aimed to analyze patterns and predictors/outcomes related to this high-risk manifestation.
Methods
This population-based analysis utilized the Surveillance, Epidemiology, and End Results (SEER) registry (2000 to 2015) to identify WDTC patients. Temporal trends and geographic variation in invasion rates were assessed. Logistic regression and propensity score matching were employed to identify predictors of secondary malignancy, mortality, and treatment impact on overall and thyroid cancer (TC)-specific survival.
Results
Of 131,721 WDTC patients, 1,662 (1.3%) had tracheal invasion and 976 (0.7%) had laryngeal invasion at diagnosis. Tracheal and laryngeal invasion rates declined from 3.7%–0.7% and 1.5%–0.6%, respectively, from 2000 to 2015. Compared to 98,835 noninvasive cases, patients with laryngotracheal invasion were older and more often male, Asian, and Hispanic (all P<0.001). This group had larger tumors with higher rates of nodal (N1: 61.8% vs. 15.1%) and distant metastases (M1: 9.3% vs. 0.4%). Age ≥55 years (hazard ratio [HR], 1.19; P=0.004) and metastases (HR, 1.75; P<0.001) increased TC-specific mortality, whereas the converse pattern was found for Asian race (HR, 0.63; P=0.002) and surgery (HR, 0.35; P<0.001). In rigorously matched groups to control confounding, adding radioactive iodine to surgery reduced mortality by 30% (P<0.001). However, external beam radiation and systemic therapy did not improve survival over surgery alone.
Conclusion
Laryngotracheal invasion is present in 0.7% to 1.3% of cases, conferring over double the mortality risk. Radioactive iodine with surgery improves outcomes in this aggressive WDTC subset.
INTRODUCTION
Thyroid cancer is the fastest-growing cancer diagnosis worldwide, marked by an alarming increase in incidence over the past few decades [1,2]. In the United States (US), it is now the fifth most common cancer among women, with more than 43,700 new cases reported annually (https://seer.cancer.gov/statfacts/html/thyro.html). Although improved imaging and diagnostic techniques have played a role in this rise, changes in tumor biology might also be contributing factors [3,4].
The prognosis for well-differentiated papillary and follicular subtypes is generally favorable, largely due to their responsiveness to standard treatments [5]. Specifically, the 10-year relative survival rate exceeds 90%, primarily due to the effectiveness of surgery and radioactive iodine therapy following diagnosis [6-8]. However, approximately 20% of patients develop regional extrathyroidal extension (ETE) and distant metastases, which significantly increase the risks of recurrence and mortality [9,10].
Of particular concern is the invasion of the thyroid capsule into adjacent laryngotracheal structures, a pathway that indicates poor differentiation and an advanced T3/T4 stage designation [11-13]. The incidence of laryngotracheal invasion is reported to range from 1% to 16% among those with well-differentiated thyroid cancers (WDTCs) [14,15]. This extensive invasion often necessitates high-risk surgeries, such as laryngectomy or tracheal resection, which are associated with significant morbidity [16,17]. Furthermore, laryngotracheal invasion is associated with a 5-year disease-specific survival rate as low as 63% and a mortality rate up to five times higher than that of noninvasive disease [15].
While detrimental to prognosis, quality of life, and survival, significant gaps remain in our comprehensive understanding of the patterns, optimal prognostication, and evidence-based management strategies for laryngotracheal invasion [18]. Many studies have limitations due to small, retrospective samples that lack diversity and generalizability [19]. In this study, we utilized the extensive Surveillance, Epidemiology, and End Results (SEER) registry, which includes data from over 130,000 US thyroid cancer patients diagnosed over nearly three decades. Our goal was to definitively analyze the demographic and clinical factors that influence outcomes throughout the disease course. It is anticipated that the findings of this study will provide insights into enhancing tools for personalized risk stratification and therapeutic decisions in cases of aggressive thyroid cancer by extensively characterizing geographic distribution, temporal trends, pathological patterns, and prognostic factors, ultimately contributing to improvements in patient care.
METHODS
Data source
Our study utilized data from the SEER program’s 17 registries, which cover approximately 28% of the US population from 2000 to 2019. The SEER database is a nationally representative, comprehensive source of population-based information, encompassing cancer incidence and survival data across various geographic areas and demographics.
Study population
The study involved a retrospective analysis of a database containing de-identified subjects, which exempted it from requiring Institutional Review Board approval (#2023-449). The initial registry included 203,728 patients diagnosed with thyroid cancer, after excluding 278 patients diagnosed incidentally at autopsy and 37 identified through death certificates. Our focus was on the 131,721 patients who had histologically confirmed cancer, comprising 123,918 with papillary thyroid cancer (PTC) and 7,803 with follicular thyroid cancer (FTC). ETE data were categorized into three groups: no ETE, confined to the thyroid (n=98,835); minimal ETE, extending to the sternothyroid muscle or perithyroid soft tissues (n=14,237); and gross ETE, involving capsular invasion, subcutaneous soft tissues, larynx, trachea, esophagus, recurrent laryngeal nerve, prevertebral fascia, major blood vessels, and mediastinum (n=18,649) [20,21]. The current study compared 2,329 patients with laryngeal (n=976) or tracheal invasion (n=1,662) to the group without ETE (n=98,835), excluding those with minimal ETE. Data of interest spanned from 2000 to 2015. Our study specifically targeted T4a tumors with laryngotracheal invasion, representing a distinct group where aggressive surgical intervention is both feasible and potentially advantageous. T4a tumors are classified as stage III disease, where aggressive surgical resection followed by adjuvant therapy is typically recommended. In contrast, T4b tumors, classified as stage IV disease, are generally deemed unresectable, with treatment shifting towards non-surgical options such as systemic therapy (e.g., tyrosine kinase inhibitors and radiation). In some instances, near-total gross resection may be considered for palliative purposes [22,23].
We identified laryngotracheal ETE using two variables from the SEER database: (1) T stage and (2) extent of invasion to adjacent structures. This was applied to 722 cases where the T stage was missing, as shown in Table 1. This method ensured the inclusion of all instances of laryngotracheal invasion in T4a tumors, capturing cases that might have been overlooked if only the T stage were considered.
Characteristics of the Study Population
Study variables
The parameters examined in our study included a broad range of demographic, clinical, and therapeutic factors. These covered age, sex, racial background, Hispanic/Latino ethnicity, urban or rural residency, household economic status, histopathological subtype, tumor size, and tumor-node-metastasis (TNM) staging. We also assessed factors such as tumor laterality, surgical interventions targeting the cancer, radiation therapy, and systemic therapeutic approaches.
Study outcomes
Our study outcomes were multifaceted, reflecting the aim to provide a comprehensive understanding of WDTC with laryngotracheal invasion. Initially, we conducted a longitudinal analysis to assess trends in the incidence of laryngotracheal invasion in WDTC, identifying any annual increases or decreases over time. Second, we compared demographic and clinical characteristics between patients with and without laryngotracheal invasion. Third, we focused on prognostic indicators by analyzing disease recurrence rates, the prevalence of second primary malignancies, overall survival, and thyroid cancer-specific survival rates across different groups. We defined a second primary malignancy as any new primary cancer that occurs in a patient with a history of thyroid cancer, using the SEER multiple primary codes and confirming through a review of site-specific factors.
We also identified independent predictors of second malignancies and mortality. Fourth, we examined the therapeutic modalities employed, including the type of surgery, use of adjuvant radiation, radioactive iodine, and systemic therapy. We analyzed the impact of treatment decisions and delays on patient survival outcomes. Fifth, we accounted for potential confounding using propensity score matching to compare survival between treatment groups, specifically for surgery alone versus surgery combined with radiation or systemic therapy. Lastly, we reported on common non-cancer causes of mortality among laryngotracheal invasion patients over long-term follow-up.
Propensity score matching
To account for potential confounding factors, we implemented 1:1 nearest neighbor propensity score matching without replacement. This method was based on a range of covariates, including age, sex, race, ethnicity, income, residence, tumor size, nodal stage, metastases, and histology. The matching process involved the following treatment groups: (1) 25 patients receiving no management were matched to 25 undergoing only surgery; (2) 345 patients treated with surgery and radioactive iodine were matched to 345 receiving surgery alone; (3) 253 patients receiving external beam radiation (EBR) in addition to surgery were matched to 253 undergoing surgery alone; (4) 48 patients treated with systemic therapy and surgery were matched to 48 undergoing surgery alone; (5) 346 patients receiving triple therapy (surgery, radioactive iodine, and systemic therapy) were matched to 346 undergoing surgery alone; and (6) 64 patients treated with systemic therapy, surgery, and EBR were matched to 64 undergoing surgery alone.
Statistical analysis
All analyses were conducted using R version 4.2.2 (R Foundation for Statistical Computing, Vienna, Austria; 2022) on macOS Ventura 13.3.1 (Apple Inc., Cupertino, CA, USA) and SPSS version 27.0 (IBM Corp., Armonk, NY, USA). Categorical variables were summarized as frequencies and percentages, while continuous variables were reported as mean with standard deviation or median with interquartile range (IQR). The chi-square and Fisher exact tests were used to compare groups for categorical data. For continuous variables, the Student t test or Mann-Whitney U test was applied, with a two-sided significance level set at 0.05. Trends in the incidence of laryngotracheal invasion over time were assessed using joinpoint regression. Logistic regression was used to identify independent predictors of recurrence and second malignancy, adjusting for confounders; odds ratio (OR) and 95% confidence interval (CI) were reported. Kaplan-Meier curves were drawn to depict survival outcomes, and the log-rank test with Bonferroni adjustment for multiple comparisons was used to assess survival differences between groups. Cox proportional hazards regression was employed to identify prognostic factors for survival, presenting hazard ratio (HR) and 95% CI.
RESULTS
Overall analysis of the study population
The workflow of patient selection is demonstrated in Fig. 1. The study examined data on 131,721 patients diagnosed with WDTC from the SEER database. Among these, 123,918 had PTC and 7,803 were diagnosed with FTC. Between 2000 and 2015, tracheal invasion occurred in 1,662 cases, while laryngeal invasion was noted in 976 cases, accounting for 1.30% and 0.70% of all WDTC patients, respectively. Notably, 309 of these patients experienced both laryngeal and tracheal invasions (Fig. 2A).
Patient selection flowchart. From an initial pool of 203,728 thyroid cancer patients in the Surveillance, Epidemiology, and End Results (SEER) database, we excluded 315 cases (278 diagnosed at autopsy, 37 from death certificates). Of the remaining 131,721 histologically confirmed cases (123,918 papillary thyroid cancer [PTC] and 7,803 follicular thyroid cancer [FTC]), patients were categorized by extrathyroidal extension (ETE): 98,835 without ETE, 14,237 with minimal ETE (excluded), and 18,649 with gross ETE. Among gross ETE cases, 2,329 had laryngotracheal invasion (976 laryngeal, 1,662 tracheal, and 309 both). The final study population comprised 101,164 patients: 98,835 without ETE and 2,329 with laryngotracheal invasion. The image was created using BioRender (www.biorender.com).
Temporal trends and patterns in the incidence and invasion characteristics of well-differentiated thyroid cancer (WDTC), papillary thyroid cancer (PTC), and follicular thyroid cancer (FTC) in the United States over 15 years. (A) A bar chart representing the total study population of 131,721 WDTC patients, of whom 123,918 had PTC and 7,803 had FTC. Of those with extrathyroidal extension (ETE), 2,329 cases presented with laryngotracheal invasion. These included 1,662 patients with tracheal invasion and 976 cases with laryngeal invasion. (B) Age-adjusted incidence rates of thyroid cancer per 100,000 people. The line graph shows a consistent increase in the age-adjusted incidence rate of WDTC over time from 2000 to 2015, demonstrating. (C) Trends of laryngotracheal invasion in WDTC patients (n=131,721). Two-line graphs overlaid on each other represent the declining rates of tracheal and laryngeal invasion in WDTC patients over the years. (D) Prevalence of tracheal and laryngeal invasion in well-differentiated thyroid cancer. A bar chart illustrating the prevalence of tracheal and laryngeal invasion in PTC and FTC patients. The bars are categorized by cancer type and invasion location. (E) A range plot detailing the yearly incidence of tracheal invasion in PTC and FTC patients from 2000 to 2015, showing a decreasing trend over time. (F) Similar to Fig. 1E, this range plot shows the yearly incidence of laryngeal invasion in PTC and FTC patients, demonstrating a declining trend.
Temporal analysis of laryngotracheal invasion in WDTC cases
The age-adjusted incidence rate of thyroid cancer consistently increased over time. In 2000, the rate was 7.67 new cases per 100,000 people. It doubled by 2015, reaching 14.89 new cases per 100,000 people (Fig. 2B). Conversely, the rates of tracheal and laryngeal invasion showed a steady decline during the same period. The incidence of tracheal invasion peaked at 3.7% in 2000 but gradually decreased to 0.7% by 2015 (P<0.001). Laryngeal invasion followed a similar pattern, starting at 1.5% in 2000 and decreasing to 0.6% by the end of the study period (P<0.001), as illustrated in Fig. 2C.
When classifying the data by cancer type, tracheal invasion was observed in 1.3% of PTC patients (n=1,587) and 1% of FTC patients (n=75). Laryngeal invasion was less common, occurring in 0.7% of PTC patients (n=928) and 0.6% of FTC patients (n=48) (Fig. 2D). The incidence of tracheal invasion in PTC patients peaked in 2000 at 3.80% and then progressively declined, reaching 0.70% by 2015. A similar decreasing trend was noted in FTC, with the highest incidence at 3.30% in 2001, decreasing to the lowest point of 0.20% by 2009, despite some fluctuations in the intervening years (Fig. 2E). Laryngeal invasion in both PTC and FTC also showed an overall downward trend. The incidence rate of laryngeal invasion in PTC decreased from 1.50% in 2000 to 0.60% in 2015. Notably, FTC exhibited a more pronounced decline, from 2.20% in 2000 to no reported cases of laryngeal invasion in 2015 (Fig. 2F).
Spatial analysis for laryngotracheal invasion in WDTC cases
The geospatial analysis of laryngotracheal invasion in WDTC patients across the US revealed significant disparities among different states, as shown in Supplemental Fig. S1. Alaska reported the highest incidence, with laryngotracheal invasion occurring in 5.7% of the diagnosed WDTC patients (n=87) within the state. In contrast, Connecticut recorded the lowest incidence, with only 1.4% of the diagnosed patients in the area (n=6,313) experiencing this complication. Additionally, California, despite having the highest number of cases (n=39,731), reported an invasion rate of 3.0%.
Demographic and clinical characteristics of patients presenting with laryngotracheal invasion
Table 1 presents an analysis of 2,329 patients diagnosed with thyroid cancer who exhibited laryngotracheal invasion, comparing them to a control group of 98,835 patients without ETE. The study revealed distinct demographic characteristics in patients with laryngotracheal invasion. These patients were more likely to be over 55 years old (59% vs. 35.5%, P<0.001), male (35.4% vs. 21.5%, P<0.001), of Asian or Pacific Islander descent (14.9% vs. 9.6%, P<0.001), Hispanic/Latino (23.7% vs. 13.9%, P<0.001), widowed (12.7% vs. 4.9%, P<0.001), and to have a history of cancer prior to the diagnosis of thyroid cancer (15.4% vs. 11.5%, P<0.001). However, there were no significant differences in residency status (P=0.51) or household income (P= 0.56) between the two groups.
In the context of pathological presentation, the classical variant of PTC exhibited a higher incidence of laryngotracheal invasion (3.80%) at the time of cancer diagnosis compared to its follicular variant (1.40%). Similarly, in FTC cases, the widely invasive variant demonstrated a higher prevalence of laryngotracheal invasion (2.50%) than the minimally invasive variant (0.60%). Intriguingly, patients with laryngotracheal invasion showed a markedly higher incidence of T4a tumors (85.4%) followed by T4b tumors (14.6%). These patients also had significantly higher rates of N1 (61.8% vs. 15.1%, P<0.001) and M1 staging (9.3% vs. 0.4%, P<0.001). Additionally, a strikingly high rate of coexisting esophageal invasion (35% vs. 0%) and distant metastasis to the lungs (10.7% vs. 0.2%) was identified in these patients.
Therapeutic strategies
Table 2 compares clinical outcomes and treatment strategies for patients with laryngotracheal invasion versus those without ETE. Regarding surgical management, although a high percentage of patients in both groups underwent cancer-directed surgery, those with laryngotracheal invasion were slightly less likely to receive this treatment compared to those without ETE (95.1% vs. 98.8%, P<0.001). Total thyroidectomy was the predominant surgical approach in both groups, yet it occurred more frequently among patients with laryngotracheal invasion (81.2% vs. 75.7%). In terms of additional therapies, patients with laryngotracheal invasion more commonly received radiation therapy (75.1% vs. 43.2%, P<0.001) and systemic therapy (54.6% vs. 47.2%, P<0.001) than those without ETE.
Clinical Outcomes and Management Strategies Employed for Patients with Laryngotracheal Invasion Compared to Those without ETE
Clinical outcomes
Clinical outcomes were significantly worse for patients with laryngotracheal invasion compared to those without ETE. Although the recurrence rates were similar between the groups (0.6% with invasion vs. 0.4% without, P=0.07), the incidence of second primary cancers was markedly higher in patients with invasion (26.7% vs. 20.6%, P<0.001) (Table 2).
Predictors of second primary malignancies
Multivariate binary logistic regression analysis identified advanced age and non-Hispanic ethnicity as significant predictors of an increased likelihood of developing second primary malignancies in patients with laryngotracheal invasion (Supplemental Table S1). Specifically, patients aged 55 years or older were more than twice as likely to develop another malignancy compared to their younger counterparts (OR, 2.078; 95% CI, 1.551 to 2.785; P<0.001). Conversely, Hispanic patients exhibited lower odds of developing additional malignancies than non-Hispanics (OR, 0.605; 95% CI, 0.433 to 0.846; P=0.003). Other variables assessed, such as sex, race, income level, residence location, tumor histology, treatment modalities, and cancer stage, did not show significant predictive value.
Survival outcomes
Among 101,164 thyroid cancer patients included in the analysis, 10,794 (10.7%) died over a median follow-up period of 14.25 years (IQR, 12.88 to 15.61). Of these deceased patients, thyroid cancer was explicitly identified as the cause of death in 1,716 (15.8%) cases. Mortality rates were significantly higher in patients with laryngotracheal invasion compared to those without ETE. Specifically, the mortality rate for patients with laryngotracheal invasion was 44.0%, versus only 9.9% for patients without such invasion (P<0.001). Additionally, 25.8% of the deaths in the laryngotracheal invasion group were directly attributed to thyroid cancer, compared to just 1.1% in the noninvasive group (P<0.001) (Table 2).
Regarding survival outcomes, the estimated 5-year overall survival rate was 96% among patients without ETE, which slightly decreased to 91% at 10 years post-diagnosis. In contrast, for those with laryngotracheal invasion, the 5-year survival rate was 75%, which dramatically declined to 60% at 10 years. Kaplan-Meier survival curves illustrated that the mean overall survival time was 216.3±0.22 months in the noninvasive group, compared to 149.9±2.04 months for those with invasion (P<0.001). Similarly, the mean thyroid cancer-specific survival was 236.1±0.09 and 172.1±2.24 months in the two groups, respectively (P<0.001) (Fig. 3).
Kaplan-Meier survival curves comparing patients with and without extrathyroidal extension (ETE). (A) Overall survival. The mean overall survival time was 216.3±0.22 months in the noninvasive group versus 149.9±2.04 months for those with laryngotracheal invasion (P<0.001). (B) Thyroid cancer-specific survival. The mean thyroid cancer-specific survival was 236.1±0.09 and 172.1±2.24 months in the two groups, respectively (P<0.001). The log-rank test was used for comparison.
Causes of death in patients with laryngotracheal invasion
Among the 2,329 patients with laryngotracheal invasion, 602 (25.8%) died from thyroid cancer during the follow-up period, while 423 (18.2%) died from other causes. The primary causes of non-cancer-related deaths included other malignant cancers (n=134 out of 423, 31.68%), heart diseases (n=93, 21.99%), cerebrovascular events (n=23, 5.44%), chronic lower respiratory diseases (n=16, 3.78%), and diabetes (n=14, 3.31%) (Supplemental Fig. S2). At the initial diagnosis of thyroid cancer, the median age for those with laryngotracheal invasion was 59 years (IQR, 46 to 71). For those who died during the follow-up, the median age at death was significantly higher, at 73.3 years (IQR, 58.9 to 86.6).
Independent predictors of mortality in laryngotracheal invasion patients
Table 3 outlines the key demographic, clinical, and treatment factors that independently predict mortality outcomes in the thyroid cancer cohort with laryngotracheal invasion. Notably, being over the age of 55 emerged as the most significant risk factor, increasing the hazard of overall mortality by 18% (HR, 1.18; 95% CI, 1.05 to 1.33; P=0.005) and cancer-specific mortality by nearly six times (HR, 5.92; 95% CI, 4.69 to 7.48; P<0.001) compared to younger patients. Additionally, male sex slightly increased the hazard of cancer-specific mortality by 29% compared to females (HR, 1.29; 95% CI, 1.08 to 1.54; P=0.005).
Predictors of Mortality in Thyroid Cancer Patients with Laryngotracheal Invasion
In terms of race/ethnicity, individuals identified as Asian/Pacific Islander experienced a 38% lower hazard of cancer mortality than Caucasian individuals (HR, 0.62; 95% CI, 0.48 to 0.84; P=0.001). Conversely, Hispanic ethnicity was associated with a 32% increase in the hazard of overall mortality, though it did not affect cancer-specific mortality (HR, 1.32; 95% CI, 1.16 to 1.52; P<0.001). Neither annual household income nor urban versus rural residence significantly influenced mortality outcomes.
T4b tumors were associated with higher overall mortality (HR, 1.38; 95% CI, 1.14 to 1.67; P=0.001) and cancer-specific mortality (HR, 1.62; 95% CI, 1.28 to 2.05; P<0.001) than T4a tumors. Consistent with the biological behavior of the tumors, the presence of nodal metastases increased the hazard by 44% (HR, 1.44; 95% CI, 1.22 to 1.78; P<0.001), and distant metastases more than doubled the mortality hazard (HR, 2.52; 95% CI, 2.04 to 3.21; P<0.001).
Undergoing primary tumor resection (HR, 0.34; 95% CI, 0.26 to 0.52; P<0.001) and radioactive iodine therapy (HR, 0.50; 95% CI, 0.41 to 0.65; P<0.001) each independently conferred a reduction in cancer-specific mortality hazard by approximately 50% to 66%. EBR and systemic therapy, however, did not significantly affect mortality outcomes in this analysis.
Impact of treatment strategies on survival
As outlined in Table 4 and depicted in Kaplan-Meier plots (Fig. 4), the choice of adjuvant treatments significantly influenced both overall and cancer-specific mortality in patients with laryngotracheal invasive thyroid cancer. The addition of radioactive iodine ablation to primary surgery significantly increased mean overall survival by 17% (167.7 months vs. 133.2 months) and cancer-specific survival by 21% (189.8 months vs. 157.1 months) compared to surgery alone (both P<0.001). In contrast, EBR did not enhance either overall or cancer-specific survival when compared to surgery alone. Similarly, systemic therapies following surgery showed no survival advantage over surgery alone, and their combination with radioactive iodine did not further improve the benefits achieved with radioiodine and surgery (P=1.0). This suggests that surgery combined with radioactive iodine provides sufficient and optimal survival outcomes. Among non-surgical options, declining all treatments significantly reduced survival (both P<0.001), highlighting the importance of definitive local-regional treatment. Regarding treatment timing, initiating therapy more than 1 month after diagnosis did not adversely affect multivariate-adjusted survival compared to earlier initiation (P>0.05).
Survival Times of Thyroid Cancer Patients with Laryngotracheal Invasion
Kaplan-Meier survival curves comparing patients who received different treatment modalities. (A) Overall survival. The curve depicts the probability of survival over time for all causes of death, comparing the different treatment modalities. (B) Thyroid cancer-specific survival. The curve shows the probability of survival without dying from thyroid cancer, highlighting the effectiveness of treatment modalities specifically against the cause of death of thyroid cancer. The log-rank test was used for comparison with the Bonferroni adjustment method. AS, active surveillance; SR, surgery+radioactive iodine ablation; SSB, surgery+systemic therapy+external beam radiation; Su, cancer-directed surgery only; SB, surgery+external beam radiation; SS, surgery+systemic therapy; SSR, surgery+systemic therapy+radioactive iodine ablation.
Propensity-matched analyses reveal survival benefit differs by radiation modality
The characteristics of patients after propensity score matching are demonstrated in Supplemental Tables S2-S7. After matching, survival was compared between treatment pairs. The combination of radioactive iodine with surgery significantly reduced mortality hazard by 30% overall (HR, 0.70; 95% CI, 0.56 to 0.86; P<0.001) and 33% for cancer-specific mortality (HR, 0.67; 95% CI, 0.50 to 0.89; P=0.007) versus surgery alone. However, EBR with surgery showed no survival advantage over surgery, with non-significant HRs of 1.11 (95% CI, 0.89 to 1.39; P=0.36) for overall mortality and 1.21 (95% CI, 0.91 to 1.61; P=0.20) for cancer death. Systemic therapy combined with surgery also failed to improve survival, with non-significant HRs of 0.81 (95% CI, 0.47 to 1.41; P=0.46) and 0.60 (95% CI, 0.29 to 1.26; P=0.18) for overall and disease-specific metrics. Moreover, triple therapy regimens demonstrated diverging survival patterns by radiation type. Adding RAI reduced the mortality hazard by 31% overall (HR, 0.69; 95% CI, 0.53 to 0.90; P=0.006) and 38% for cancer-specific mortality (HR, 0.62; 95% CI, 0.43 to 0.88; P=0.009). However, substituting EBR raised the hazard by 58% (HR, 1.58; 95% CI, 1.01 to 2.47; P=0.046) and 108% (HR, 2.08; 95% CI, 1.17 to 3.71; P=0.013) (Table 5).
Hazard Ratios for Mortality Comparing Treatment Groups after Propensity Score Matching
DISCUSSION
Our extensive analysis of more than 130,000 thyroid cancer patients revealed a declining trend in the incidence of laryngotracheal invasion, yet there remained a persistent, disproportionate mortality burden. In addition to uncovering dynamic geospatial and longitudinal patterns, we identified a comprehensive range of prognostic indicators and demonstrated significant survival benefits associated with the selective use of adjuvant radioactive iodine after surgery for high-risk invasive disease.
Despite thyroid cancer being identified as the most rapidly increasing malignancy in the US [2], our data clearly shows that the rates of tracheal and laryngeal invasion have decreased by over 50% since their peak in 2000. However, with 1,662 (1.3%) tracheal and 976 (0.7%) laryngeal invasive cases recorded among our 131,721 patients from 2000 to 2015, this phenotype remains clinically significant. This is due to its association with more than double the mortality risk compared to noninvasive disease and a 5-year survival rate below 60%, indicating a persistently poor prognosis. The decline in incidence rates likely reflects the delayed effects of increased screening and earlier detection at premalignant stages [24-26].
We also discovered substantial state-level variation in invasion, with rates of 1.4% to 5.7%. This likely reflects the impact of factors such as access to diagnostics and care, socioeconomic status, and environmental triggers on the detection of occult invasion and subsequent outcomes [27-29]. Our data provide a basis for reallocating resources to enhance early-stage diagnosis in demographics at higher risk. Further investigation into facility-level contributors could reveal actionable disparities and inform health policy initiatives aimed at targeting the highest-risk demographics [30].
Our regression models demonstrated prognostic utility for guiding frontline risk stratification. Factors such as advanced age, male sex, Hispanic ethnicity, nodal involvement, and distant metastases independently predicted the highest mortality risks, even after adjusting for confounders. Tailoring screening and management strategies based on these factors could enhance triage to specialty care and reduce adverse outcomes [30]. Interestingly, income and urban residency had negligible predictive value when controlling for nodal involvement and metastatic spread. This indicates that socioeconomic barriers to accessing specialty oncologic care might be overcome in the current healthcare environment [31]. Additionally, a 37% reduction in cancer-specific mortality among Asian patients has emerged as a novel, previously unrecognized protective factor. This finding warrants further investigation to determine whether environmental, genetic, or lifestyle factors are involved [32,33]. Elucidating the underlying factors in these ethnicities could significantly shape preventative and therapeutic approaches [34].
However, the most important finding from our up-to-date dataset and rigorous propensity-matched analysis is the significant 30% survival advantage derived solely from supplemental radioactive iodine following surgery for invasive disease. While previous studies were limited by small sample sizes and lacked appropriate comparison groups to definitively demonstrate the benefits of radioactive iodine [35-37], our latest evidence addresses these gaps. Given the high 10-year mortality rate, which exceeds 25%, this 30% reduction in hazard with radioactive iodine prevents early deaths during the initial high-risk years when locoregional progression peaks, before the survival curves gradually converge. This intervention also results in meaningful gains in quality-adjusted life years, as previously noted [38]. In our real-world data, neither EBR therapies nor systemic agents provided additional benefits beyond those achieved with radioactive iodine and surgery, unlike findings from earlier studies [39,40]. Collectively, these latest data strongly endorse updated guidelines [41] recommending surgery with radioactive iodine remnants ablation as the optimized evidence-based standard of care for regionally invasive thyroid cancer.
Despite the extensive contemporary dataset, several limitations must be acknowledged. Residual confounding due to unmeasured variables, the retrospective nature of the SEER data, and potential coding errors are inherent limitations. The SEER database, being a population-based registry, may exhibit different reporting patterns for ETE compared to institution-based studies, which could affect the proportion of cases without ETE in our analysis. Additionally, the database lacks detailed information on certain clinical variables, which limits our ability to conduct more comprehensive multivariate analyses. The generalizability to less common histological variants requires further investigation. Nonetheless, our multi-pronged analysis of population-based data, covering 16 years and involving over 130,000 patients, provides a robust assessment of this rare outcome. Future studies that include emerging mutational, transcriptomic, and radiographic predictors could further clarify which subgroups might benefit from adjuvant systemic therapies or EBR.
In conclusion, our contemporary analysis of over 130,000 thyroid cancer patients highlights a decline in laryngotracheal invasion, focusing on 2,329 residual cases of papillary and follicular types. In addition to identifying diverse prognostic indicators, we found a significant 30% survival benefit when radioactive iodine was added to surgical treatment for this invasive disease, potentially reducing early mortality in this high-risk group. We aim for our real-world data to inform risk-adjusted guidelines that optimize the selection of adjuvant therapies and enhance patient outcomes. Advancing meaningful progress in the treatment of rare and aggressive cancer manifestations requires a thorough examination of uncommon histological types that are often neglected due to a gradual increase in less aggressive subtypes. Equally important is the need for future research to explore the generalizability and biological mechanisms of these findings. Continuing to translate population-level analytics into clinical practice to better manage these high-risk variants remains a top priority.
Supplementary Material
Supplemental Table S1.
Predictor Risk Factors for Second Primary Malignancies in Thyroid Cancer Patients with Laryngotracheal Invasion
Supplemental Table S2.
Characteristics of Matched Cohorts in Surgery and RAI versus Surgery
Supplemental Table S3.
Characteristics of Matched Cohort Treated with External Beam Radiation Plus Surgery Matched to Surgery Alone
Supplemental Table S4.
Characteristics of Matched Cohorts in Systemic Therapy and Surgery versus Surgery Alone
Supplemental Table S5.
Characteristics of Matched Cohorts in Triple Therapy versus Surgery Alone
Supplemental Table S6.
Characteristics of Matched Cohorts in systemic Therapy+Surgery+EBR versus Surgery Alone
Supplemental Table S7.
Characteristics of Matched Cohorts in Surgery versus No Management
Supplemental Fig. S1.
Geographical variation in the percentage of tracheolaryngeal invasion among well-differentiated thyroid cancer patients. The bar chart depicts the geographical variation in the prevalence of tracheal and laryngeal invasion in thyroid cancer patients across different states in the United States from 2000 to 2015. Each bar represents a specific state, with the height indicating the percentage of patients in that state who experienced tracheolaryngeal invasion at the time of their cancer diagnosis. The number of patients diagnosed in each state during the study period is indicated in parentheses next to the state name.
Supplemental Fig. S2.
Prevalence of top non-thyroid causes of mortality in patients with laryngotracheal invasion. Data are presented as percentages.
Notes
CONFLICTS OF INTEREST
No potential conflict of interest relevant to this article was reported.
ACKNOWLEDGMENTS
This study was funded by the ThyCa: Thyroid Cancer Survivors’ Association Inc. and administered by the American Thyroid Association through grant number (THYROIDGRANT 2021-0000000232) and The School of Medicine Pilot Grant (to Eman A. Toraih). We acknowledge the Surveillance, Epidemiology, and End Results (SEER) Program team for establishing and maintaining this valuable cancer database, allowing robust epidemiological research on thyroid cancer prognosis.
This work was supported by the Tulane Cancer Center, part of Tulane School of Medicine, and a consortium partner of the Louisiana Cancer Research Center
AUTHOR CONTRIBUTIONS
Conception or design: E.A.T., J.A.J., M.H.H., A.A.M.S., M.S.F., E.K. Acquisition, analysis, or interpretation of data: E.A.T., J.A.J., M.H.H. Drafting the work or revising: E.A.T., J.A.J., M.H.H., A.A.M.S., M.S.F., E.K. Final approval of the manuscript: E.A.T., J.A.J., M.H.H., A.A.M.S., M.S.F., E.K.