Imaging evaluation of lung cancer treated with PD-1/PD-L1 inhibitors

Abstract

Immunotherapy (PD-1/PD-L1 inhibitors) has attracted attention for lung cancer treatment and recasted the administration of immunotherapeutics to patients who have advanced/metastatic diseases. Whether in combination or as monotherapy, these medications have become common therapies for certain patients with lung cancer. Moreover, their usage is expected to expand widely in the future. This review aims to discuss the imaging evaluation of lung cancer response to PD-1/PD-L1 therapy with focus on new radiological criteria for immunotherapy response. Abnormal radiological responses (pseudoprogression, dissociative responses, and hyperprogression) and immune-related adverse events are also described.

Introduction

Lung cancer (LC) is a disease characterized by a firmly immunosuppressive environment. Under this assumption, various types of immunization have undergone clinical research but only achieved minimal success. Initial findings have prompted tremendous interest in immune checkpoint inhibitors (ICIs).¹ Programmed cell death protein 1 (PD-1) inhibitors are the mainly studied agent and have shown remarkable efficacy in trials of non-small-cell lung cancer (NSCLC).^2–4 In particular, consolidation durvalumab programmed cell death ligand 1 (PD-L1) after chemoradiotherapy for Stage III NSCLC is the standard approved therapy.

The PD-1 inhibitors pembrolizumab and nivolumab and PD-L1 inhibitor atezolizumab have been recently endorsed by the United States Food and Drug Administration as NSCLC treatment and are currently approved for the second-line treatment of patients with advanced LC. These medications have improving effects in terms of increased endurance, target reaction rate through randomized CTs, and superior toxicity profile to comparator (docetaxel).⁵ However, the immune checkpoint blocking by PD-1/PD-L1 is linked to a set of side-effects known as immune-related adverse events (irAEs), which impact various organs across the body.^6,7

Imaging has become a crucial part of evaluating treatment reactions for patients treated with ICI and identifying irAEs that may include different organs.^6,7 For efficient clinical decisions and patient management, timely irAE diagnosis via imaging is critical.^6,8 With the growing usage of ICI for LC in clinical settings, radiologists must be aware of the abnormal tumor responses and widespread adverse events with imaging in patients receiving immunotherapy.

This review aims to discuss the imaging evaluation of LC response to PD-1/PD-L1 therapy with focus on the new radiological criteria for immunotherapy response. Abnormal radiological responses (pseudoprogression, dissociative responses, and hyperprogression) and irAEs are also described.

The rest of this paper is organized as follows. Section two explains the radiological response criteria for the assessment of lung cancer immunotherapy treatment. Section three discusses the abnormal radiological reaction patterns. Section four presents details on irAE imaging, and section five provides the conclusion.

Assessment of immunotherapy treatment response

A number of radiological standards of response to various cancer therapies have been published in recent decades.⁹ The World Health Organization (WHO) or solid tumor response assessment (RECIST) initially relied on cytotoxic chemotherapy and presumed that the increase in tumor size or new lesion emergence indicated progressive disease (PD); as a result, therapy is stopped.^10,11

Immunotherapy for LC has recently gained popularity. ICIs are currently widely used in various clinical settings, and their utilization is predicted to grow in the future. However, their specific mechanism can produce abnormal response patterns on imaging that cannot be accurately assessed using traditional response standards such as RECIST and WHO. Therefore, multiple response criteria [“immune-related response criteria (irRC), immune-related RECIST (irRECIST), immune RECIST (iRECIST), and immune-modified RECIST (imRECIST)”] have been suggested and used in immunotherapy clinical studies over the last few years.¹²

This section discusses the current response standards for immunotherapy that share some characteristics, such as evaluating the target and non-target lesions in baseline studies. Overall response is assessed by recording the occurrence of new lesions, measuring changes in the size of target lesions, and estimating variations in non-target lesions. This parameter is classified into four response categories [stable disease (SD), PD, partial response (PR), and complete response (CR)]. The main differences between each response criterion of immunotherapy lie in how measurable lesions are measured (1D or 2D based on RECIST or WHO, respectively), how new lesions are defined within the PD category, and whether new lesions are added to the index tumor burden.⁵ Table 1 shows the comparison among RECIST 1.1, irRC, irRECIST, iRECIST, and imRECIST criteria.

Table 1.

Comparison of solid tumor response assessments (RECIST)

	RECIST 1.1	irRC	irRECIST	iRECIST	imRECIST
CR	Complete disappearance of all lymph nodes or target lesions < 10 mm in the short axis	Complete disappearance of all target lesions and no new lesions (should be confirmed by two consecutive findings separated by at least 4 weeks)	Complete disappearance of all lymph nodes or target lesions < 10 mm in the short axis, no new lesions, response confirmation is not required	Complete disappearance of all lymph nodes or target lesions < 10 mm in the short axis, no new lesions	Disappearance of all lesions (should be confirmed 4 weeks after the initial disappearance is discovered)
PR	≥30% decrease in tumor burden or ≥15% decrease in tumor attenuation at CT, no new lesions	≥50% decrease in tumor burden in comparison to baseline (should be confirmed by two successive findings separated by 4 weeks or more)	≥30% decrease in TMTB in comparison to baseline, no new lesions	≥30% decrease in tumor burden or ≥15% decrease in tumor attenuation at CT, no new lesions	≥30% decrease in tumor burden from the baseline (if there are no new lesions or if non-target lesions progress)
PD	≥20% increase in SLD of target lesions (absolute increase ≥5 mm) relative to nadir study, or the emergence of new lesions.	≥25% increase in SLD with respect to nadir (should be confirmed by two successive findings separated by 4 weeks or more)	≥20% increase in TMTB with respect to nadir study (absolute increase ≥5 mm)	iUPD: ≥20% increase in tumor burden with respect to nadir or the emergence of new lesions confirmation is required between 4 and 8 weeks after the first iUPD is discovered. iCPD: iUPD is confirmed as ICPD, only if further new lesions emerge in consecutive CSI, or if the size of previous new lesions increases by at least 20% (≥5 mm for the sum of new target lesions or any increase in new non-target lesions)	PD is identified only from measurable lesions. If a “no progression” state is achieved, then progression is not confirmed (4 weeks after PD is discovered). If SLD (RECIST) is ≥20 %, PD is reached in comparison to nadir.
SD	None of the above	None of the above	None of the above	None of the above	None of the above
New Lesions	Results in PD	New lesions are added to TTB, the emergence of new lesions does not immediately indicate PD.	New lesions are added to the TMTB, the emergence of new lesions does not immediately indicate PD. unambiguous progression of new non-measurable lesions indicates the presence of PD (irPD)	Results in iUPD and thus in iCPD when new lesions further emerge or the size of new lesions increase (≥5 mm for SLD or any increase in the number of non-target lesions)	Emergence of new measurable lesions do not immediately indicate PD and are added to TTB. If new non-measurable lesions appear, then this condition does not indicate PD but rules out CR
PD Confirmation	Not required	Required	Recommended to be at least 4 weeks after discovering irPD for the first time	Required	Required
Consideration of clinical status	Not specifically mentioned	Not specifically mentioned	Not specifically mentioned	Taken into consideration when making a decision about or not to continue therapy after iUPD	Not specifically mentioned

CR, Complete response; CSI, Cross-sectional imaging; PD, Progressive disease; PR, Partial response; SD, Stable disease; SLD, Sum of largest diameters; TMTB, Total measured tumor burden; TTB, Total tumor burden; iCPD, Immune confirmed progressive disease; iPR, Partial immune response; iUPD, Immune unconfirmed progressive disease; irPD, Immune-related progressive disease.

RECIST 1.1

These standards were launched in 2009 and simplified the previous therapeutic evaluation version. The RECIST standards specify the minimum size of target lesions (two lesions per organ, maximum of five) and the use of a 1D measure. Within these standards, PD is defined as the emergence of new measurable lesions or an increase of 20% or more in SLD [“the sum of the longest diameters of target lesions” (absolute increase ≥5 mm) compared with nadir (best response)].¹² These criteria are the most widely used for “therapeutic” response evaluation in solid tumors and are also the only standards approved by considerable regulatory authorities.^10,13,14

irRC

These standards were launched in 2009. These criteria are based on a modification of WHO standards and are substantially different from RECIST 1.1. The number of target lesions to be assessed (5 per organ, up to 10 visceral) is greater than that in RECIST 1.1. irRC criteria use 2D rather than 1D measure. Moreover, the emergence of new lesions does not indicate PD but are added to the total tumor burden (TTB). If SLD is ≥25% relative to nadir, then PD is reached. Any PD should be confirmed radiologically within 4 weeks.^15,16

Despite the above-mentioned advancements, irRC criteria have the following limitations: (1) based on the WHO standard, and thus cannot be compared with the RECIST standard when utilizing 2D measurements, (2) provide less information on adenopathy diseases, and (3) measuring a large number of target lesions is time consuming.^17–20 Although some literature data indicated an underestimation of pembrolizumab’s effectiveness in 15% of patients with melanoma by using classical RECIST 1.1 standards compared with irRC²¹, the data on patients with LC and receiving ICI were unclear. For example, recent studies evaluating ICI response in patients with LC indicated that irRC and RECIST 1.1 have similar prognostic importance.²²

irRECIST

In 2014, these standards were established to align irRC with RECIST 1.1 and create a reporting system that is fast and reproducible. irRECIST standards use 1D measurement and few target lesions to allow for comparability with RECIST-based studies. In addition, these criteria are comparable with RECIST 1.1 regarding the selection criteria for target lesions (two lesions per organ, maximum of five) and the manner of measuring these lesions. However, irRECIST differs from RECIST 1.1 in how newly discovered lesions are included in response evaluation. Different from that in RECIST 1.1, the emergence of new lesions does not automatically indicate PD, and new measurable lesions are added to “the total measured tumor burden (TMTB)” in irRECIST. Therefore, new lesions should be classified as measurable or non-measurable in irRECIST, and the chosen new targets should match the same inclusion criteria as the baseline lesions. If new target lesions appear, then “the longest diameters of existing non-nodal target plus new non-lymph node target lesions, and short-axis diameters of existing lymph node targets plus new lymph node target lesions constitute the TMTB”.¹² This approach allows the continuation of a possibly efficient treatment when new lesions appear.

irPR (partial response) and PD thresholds are consistent with RECIST 1.1. Although PD confirmation is not required, for patients with a slight increase in TMTB >20%, progression must be confirmed at least 4 weeks after irPD is identified for the first time.^15,16,20

iRECIST

These standards were issued in 2017 and created from an agreement between regulatory agencies, the pharmaceutical industry, and the RECIST working group to systematize and verify immunotherapy response standards. Rules were developed to guarantee consistent data gathering and to make the comparative analysis of CTs based on immunotherapy easier. iRECIST are similar to RECIST 1.1 and irRECIST in terms of 1D measurement, suggested imaging modalities, and definition of target lesions (two per organ, five in total).^23,24 However, in contrast to irRECIST, the emergence of new lesions is evaluated as per RECIST 1.1 but are separately recorded (not added to the SLD of baseline-identified target lesions).

iRECIST response categories include iSD (stable disease), iCR (full response), and iPR (partial response), iUPD (unconfirmed PD), and iCPD (confirmed PD). iUPD and iCPD categories allow for the accurate characterization of abnormal responses. In consideration of target lesions, iSD, iPR, and iCR are still feasible after iUPD appears, whereas iCPD is not detected. iUPD is defined as the emergence of new lesions or an increase by ≥20% in tumor burden relative to nadir. iUPD is confirmed as ICPD only when further new lesions emerge in consecutive cross-sectional imaging (CSI) or the size of previous new lesions increases by at least 20% (any increase in new non-target lesions or ≥5 mm for the sum of new target lesions) relative to nadir. The RECIST workgroup recommends performing the next CSI assessment in 4–8 weeks after iUPD is discovered to allow for treatment continuousness and salvage therapy when necessary.^18,25,26

imRECIST

These standards were defined in 2018. These criteria are modifications of the RECIST 1.1 system based on the principles of irRC. imRECIST criteria permit gathering additional scans and the best overall response to occur in patients who are still receiving therapy after PD radiological evaluation. These standards select two lesions per organ with a maximum of five and use 1D measurements.

The emergence of new measurable lesions does not automatically indicate PD and are added to TTB. If new non-measurable lesions appear, then this condition does not indicate PD but rules out CR. In addition, progression in non-target lesions does not imply PD. If a “no progression” state is achieved, then progression is not confirmed (4 weeks after PD is discovered). If SLD (RECIST) is ≥20 %, PD is reached in comparison to nadir. imRECIST criteria relate progression or response patterns with overall survival by using indirect assessment criteria.²¹

Abnormal radiological response patterns

Traditional response standards rely on “cytotoxic chemotherapy” and presume that the increase in the tumor size or new lesions' appearance indicates PD; as a result, therapy is stopped.^10,11 However, in immunotherapy-treated patients, CR, PR, or SD might be attained despite the emergence of new lesions or an initial increase in the tumor size (in some patients).^20,27 This abnormal response is known as pseudoprogression and could be wrongly classified as PD when traditional criteria (e.g. RECIST) were used. “Dissociated or paradoxical responses” and hyperprogressions are other examples of abnormal immunotherapy responses.^28,29

These abnormal responses must be considered in LC treatment because traditional standards may underestimate the efficacy of ICI in 11% of cases due to their failure to appropriately interpret the disconnected or paradoxical response and pseudoprogression.³⁰

Pseudoprogression

Pseudoprogression is a phenomenon in which patients receiving ICI show response or stability after an initial increase in tumor burden or in spite of the presence of additional lesions.³¹ This condition denotes false tumor progression; therefore, an initial increase in tumor burden (or the emergence of new lesions) on radiology in LC immunotherapy-treated patients should be confirmed during a specific period before assigning “real” tumor progression and discontinuing therapy (most researchers recommend that any progression should be confirmed 1–2 months after first radiological evidence).⁵ This phenomenon ordinarily occurs within the first 12 weeks of beginning immunotherapy, though late-stage symptoms have been identified.³²

Pseudoprogression appears in 0.6–5.8% of patients with LC who are being treated with immunotherapy; hence, most patients who show initial radiological progression will encounter real progression.³³ This condition is a challenge for clinical physicians and radiologists. To date, no adequate radiological or biochemical marker can help distinguish between pseudoprogression and real progression.³⁴ The most critical outcome of this phenomenon is that patients get more survival and clinical benefit than their counterparts with real progressions. In a recent study on 160 patients receiving ICI treatment, 5% showed pseudoprogression, and all had achieved a clinical benefit.³⁰ Figure 1 shows details on pseudoprogression.

Figure 1.

Pseudoprogression accompanied by an initial increase in tumor size followed by subsequent tumor shrinkage as a result of an immune-related response in a 64-year-old male with right lower lobe lung squamous carcinoma treated with pembrolizumab. The right images are the window of mediastinal, and the left images are the window of pulmonary. (a) Baseline axial CT image collected prior to treatment indicates a lower right lobe mass near the pleura (arrow). (b) Axial CT images follow-up after 2 months of treatment indicates an increase in tumor size (arrow) and pleural effusion. In addition, lytic destruction of the neighboring rib and many swollen lymph nodes in the right hilum and mediastinum are observed. (c) Other follow-up axial CT images show a remarkable reduction in mass size at 4 months of therapy (arrow). Additionally, the neighboring rib was substantially alleviated, whereas the right hilum and mediastinal lymph nodes were greatly reduced. IO, immunotherapy (reprinted with permission from Wu et al³⁵ copyright: ^© RSNA, 2019).

Dissociated or contradictory response

Contradictory or dissociated response involves a disparate tumor lesion reaction in the same patient, with shrinkage in some lesions and progression in others.²⁹ This phenomenon emerges within the first few months after starting immunotherapy but has not been studied in depth. The most crucial implication of this atypical response is that patients experience some clinical benefits and have higher rate of survival than patients who have seen their true progression. However, these advances are less than those for pseudoprogressive patients.⁵

Hyperprogression

Hyperprogression is a remarkable increase in tumor burden with inadequate diagnosis in patients receiving immunotherapy for the first time.³⁶ Progression is faster than the reported or predicted level for that cancer after immunotherapyis initiated. “This abnormal response is not unique to immunotherapy, but it is documented more frequently with this approach than with other therapies (9%–16%)”.³⁷ Different from other atypical responses, the most critical consequence of the hyper progression is poor survival, i.e. patients die in a few weeks’ time.³⁸

Imaging of irAEs

irAEs are autoimmune effects originating from the improper activation of the immune system through immunotherapy.³⁹ These events differs from cytotoxic chemotherapy. Hence, radiologists must know these adverse events and their radiological characteristics for early diagnosis and immediate treatment.⁴⁰ The most widespread irAEs are skin diseases, comprising rash, erythema, and vitiligo. The observable irAEs in radiological checks are colitis, pneumonitis, thyroiditis, arthritis, pancreatitis, and hepatitis.^41,42

With the growing availability of immunotherapies, radiologists must be aware of the imaging and clinical symptoms of irAEs and participate in optimal monitoring and early diagnosis as multidisciplinary team members for cancer care.⁴³

Pneumonitis

This disease is characterized by “focal or diffuse exudation, inflammation, and occasionally lung parenchyma fibrosis”.⁴⁴ ICI-related pneumonitis is of concern, particularly in patients who received ICI. Despite being relatively uncommon, pneumonitis is potentially life-threatening and clinically dangerous. “Pneumonitis-related deaths in early phase trials is recognized as an event of special interest”.^45–47

Pneumonitis can radiologically and clinically mimic tumor progression or lung infection in patients with LC.³² Immune-related pneumonitis has been reported more commonly in patients with advanced LC (3%–5%) than in those with malignant melanoma (0–2%).⁴⁸ The incidence of pneumonitis is higher in patients receiving pembrolizumab or nivolumab monotherapy and nivolumab/ipilimumab combination therapy than in those administered with ipilimumab monotherapy.^49,50

Immune checkpoint blockade (ICB) therapy-related pneumonitis can occur more often and more rapidly in NSCLC than in other types of tumors (particularly melanoma).^51–54 Pneumonitis is a particular concern throughout treatment schedules for unresectable Stage III NSCLC with ICB after chemoradiotherapy because the lung function may already be affected by tumor size and position, comorbid lung conditions, and underlying smoking sequelae. In some patients, radiotherapy can trigger a dose- and irradiated volume-dependent pneumonitis.⁵⁵ As stated in the American Thoracic Society/European Respiratory Society classification for interstitial pneumonia, Nishino et al⁵² presented the radiological performance of 20 patients with PD1-associated pneumonitis as follows: “65% with cryptogenic organizing pneumonia (COP), 15% with unspecific interstitial pneumonia (NSIP), 10% with hypersensitivity pneumonitis (HP), and acute respiratory distress syndrome/acute interstitial pneumonia (ARDS/AIP).” The lower lung zone is more regularly affected than the middle and higher zones.⁵² Multifocal and mixed distributions are more frequent than basal and peripheral distributions. Reticular patterns and consolidations were found in most cases, and frosted glass densities were observed in all patients.⁵² Figure 2 shows an example of a COP pattern.

Figure 2.

A 73-year-old male with Grade 3 pneumonitis after the third sequential dose of ipilimumab after nine doses of pembrolizumab due to tumor progression. (a, b) Pattern of cryptogenic organizing pneumonia. Mixed and multifocal peripheral ground glass opacities and consolidations. (c, d) Improvement in findings after stopping ipilimumab and corticosteroids. The arrows in (b, d) show multiple lung metastases (reprinted with permission from Widmann et al.⁴¹)

A few instances of pneumonitis caused by PD-1 inhibitors have been reported, and the radiographic and clinical subtleties in NSCLC have been depicted.⁵³ These cases revealed that PD-1 inhibitor-related pneumonitis might have a range of radiologic appearances that include modified radiographic examples depicted in the ATS/ERS arrangements of interstitial types of pneumonia.^47,53

Clinical treatment courses of pneumonitis are additional factors affecting patients; a few were expectedly admitted to the intubation and emergency unit, and others dealt effectively with oral corticosteroids and had the option to restart their anti-PD-1 treatment without recurrent pneumonitis.^47,53 Patients with pneumonitis due to Nivolumab treatment were treated with corticosteroids for an average duration of 6.1 weeks.⁵² Pneumonitis is diagnosed using a combination of clinical, radiological, and analytical data, which require interdisciplinary cooperation.⁵⁶

Colitis

Colitis related to immune medication is one of the most common organ-specific irAEs, particularly for ipilimumab-treated patients. Immune-related colitis rate varies from 8 to 22%.^6,57 Diffuse and segmental colitis are two distinct patterns of colitis associated with immune diseases. The former, which accounts for approximately 75% of cases, is described by a moderate diffuse thickening of the intestine wall or a fluid-filled puffed colon with mesenteric engorgement for the vessels.^6,7,58 The latter, which accounts for 25% of cases, is connected with diverticulosis segmental colitis associated with diverticulosis (SCAD) described as “associated pericolic fat stranding and mild wall thickening in the pre-existing diverticulosis segment”.^6,7,58 In addition, colitis can lead to perforation under certain extreme conditions and can be easily identified by non-contrast CT or plain radiography.⁴²

Hepatitis

Hepatitis due to immune therapy is a rare event observed in 1–2% of patients.^6,59 Its median time is 5 weeks “(range, 1–49 weeks) in patients treated with PD-1/PD-L1 or CTLA-4 inhibitor monotherapy”.⁶⁰ Hepatitis has widely common forms varying from asymptomatic light cases with mild changes in liver function to serious cases with severe symptoms and high hepatic function test results. Hepatomegaly, periportal lymphadenopathy, and periportal edema are frequently associated with severe cases.^6,7,61 In an investigation of irAEs including 39 patients treated with PD-1, hepatitis was confirmed in 2 patients with steatosis, periportal, lymphadenopathy, hepatomegaly, and ultrasound gallbladder edema.⁶² In a meta-analysis of irAEs, a high level of ALT or AST was observed in 4–5% of patients with nivolumab treatment.⁶³ By contrast, the majority of patients with hepatitis have low ALT and AST, and the diagnosis depends on CT.⁶⁴

Sarcoid-like lymphadenopathy and granulomatosis

Sarcoid event is an uncommon but serious systemic complication that can emulate PD in patients with lung cancer treated using ICI.³³ Sarcoid-like lymphadenopathy and granulomatosis often present in the form of mediastinal and hilar lymphadenopathy and are occasionally accompanied by multifocal pulmonary nodular anomalies similar to parenchymal sarcoidosis with a tendency to engage lymphatic systems.^6,65,66 This condition often has no symptoms and, with no specific treatment, can be self-limiting. Patients with symptoms can be treated with corticosteroids and achieve good response.⁶⁵ From the radiological perspective, this event is described by lymphadenopathies’ appearance and tiny perilymphatic pneumonic knobs; hence, PD is difficult to distinguish from carcinomatous lymphangitis.⁶⁷ Although granulomas are important for diagnosis confirmation, they are not easily detected; hence, the diagnosis is mostly produced by integrating radiological and clinical data after potential causes have been eliminated (such as tumor progression or infections).⁶⁸ Generally, patients with sarcoid reaction show good overall status with tumor response signs in other organs and react rapidly to corticosteroids. Compared with pneumonitis, a sarcoid reaction does not demand immunotherapy discontinuation.⁶⁹

Other organ-specific irAEs and unmet clinical needs

Numerous other organ-explicit irAEs, such as encephalitis, myelitis, joint pain, nephritis, pancreatitis, and myocarditis, have been documented. Some serious cases of unsusceptible associated myocarditis have been recently published.^70,71

Conclusion

PD-1/PD-L1 inhibitors represent a modern class of medications that have dramatically changed lung cancer treatment methods. As a result, imaging has become crucial for evaluating treatment reactions for patients treated with these inhibitors. Based on the imaging-specific demands, radiologists are more involved in treating patients who receive ICI by interpreting staging and restaging exams, and therefore must be aware of the abnormal tumor responses and widespread adverse events observed during imaging. In addition, radiologists need to have access to very detailed referrals from oncology colleagues to allow them to interpret these changes more accurately. Therefore, the wide usage of immune-checkpoint inhibitors has increased the demand for radiological competencies to clarify irAEs.