Epidemiology

Highest rates in Eastern Asia, and south and east Africa
In Australia - incidence 5.1 per 100,000 in 2023
More common in men

Pathophysiology

Think of it as two separate disease processes, with separate risk factors
- However, controversial whether this actually influences treatment approach
- SCC tends to respond more favourably to chemoradiotherapy
Check biomarkers:
- dMMR - checkpoint modulator
- PD-L1 - checkpoint modulator
- HER2 for gastric/GOJ adenocarcinoma - trastuzumab

Adenocarcinoma (60%)

- Risk factors:
  - Modifiable:
    - Smoking
    - Obesity
    - NOT alcohol
  - Non-modifiable:
    - GORD/Barrett's/oesophageal acid exposure - duration and frequency of symptoms is correlated with risk
    - There are familial forms of Barrett's that increase the risk
    - Caucasian men
- Generally found at GOJ and a/w Barrett's
- Increasing prevalence in developed countries
- Adenocarcinomas usually metastasise intra-abdominally

SCC (40%)

- Risk factors:
  - Modifiable
    - Smoking and alcohol are the main risk factors (~90% of cases)
      - Synergistic effect on risk
    - Dietary factors - N-nitroso, hot drinks
    - Caustic ingestion
  - Non-modifiable
    - HPV (probably only a small proportion of cancers)
    - Plummer-Vinson syndrome
    - Achalasia
    - Upper aerodigestive tract SCC
    - Tylosis
    - Fanconi anaemia
    - 4x more prevalent in men
- Pathophysiology
  - Premalignant lesion is squamous dysplasia (a.k.a. squamous intra-epithelial neoplasia). Over a 13.5 year follow-up, SCC developed in 24%, 50% and 74% of patients with mild, moderate and severe dysplasia at baseline, compared to 8% of patients with normal index histology.
  - Squamous dysplasia can regress and relapse spontaneously. Can be present anywhere in the oesophagus. Chromoendoscopy with Lugol's iodine or NBI helps to identify it.
  - Squamous dysplasia should be treated with EMR/ESD depending on size

Generally proximal- or mid-oesophagus - at or higher than carina
Decreasing prevalence in developed countries
Usually metastasises intra-thoracic, and metastasises earlier than adenocarcinoma
Worse prognosis

Rare malignant tumours

- Small cell carcinoma (0.6%)
  - Aggressive phenotype, similar to other poorly-differentiated neuroendocrine cancers
  - Typically present with lymph node involvement already
  - Overall prognosis poor but long-term survival is possible with curative resections
- Primary melanoma of the oesophagus (0.1%)
  - Generally late stage at presentation, with poor prognosis
- Leiomyosarcoma
  - Far less common than leiomyoma
  - Often erode through mucosa, appearing as an ulcerated or exophytic mass
  - EUS - irregular borders and more heterogenous than would be seen with leiomyoma
  - Oesophagectomy with radical lymphadenectomy is treatment of choice, although prognosis is poor
- GIST
  - Similar appearance to leiomyoma, differentiated by CD117 (c-kit) and CD34 stain positivity. Also tend to be larger than leiomyomas, with uptake of IV contrast on CT and significant PET avidity.
  - Can be enucleated provided negative margins can be obtained; if not, formal oesophagectomy will be required
  - Consider imatinib for any GIST >3cm or with high-risk factors, or as neoadjuvant therapy for locally advanced tumours
  - Lymph node metastasis is unusual
  - Worse prognosis than gastric GIST

Clinical manifestations - really only seen in advanced disease

There are two scenarios. Either presenting with dysphagia (late) or picked up on screening (earlier), but the majority are only found once symptomatic.

Progressive dysphagia - once lumen <13mm (two-thirds obstructed), generally signifying T3 disease
Weight loss - mostly due to dysphagia-related change in diet
Odynophagia seen in 20%
GIT blood loss ?anaemia
Adenocarcinoma - long history of reflux symptoms
Fatigue, retrosternal pain
Resp - aspirations, cough, hoarse voice (?RLN involvement)
Late - tracheobronchial fistulae
Check for cervical and supraclavicular lymph nodes

Diagnosis:

Gastroscopy and biopsy with forceps - multiple biopsies should be taken for any suspicious lesions
- Classically appears as friable, ulcerated masses
- Early stage - ulcerations or small nodules
- Can present as strictures or ulcerations
- Location should be noted
- Presence of Barrett's should be defined according to Prague criteria
Sometimes seen on barium swallow

Workup

CT CAP with IV contrast
- Look for mets, and locoregional extent
- Not as useful in early stage disease, and can't distinguish T3 from T4 very well
- Look for lymph nodes >1cm in short axis
FDG PET/CT
- Should be done in most cases to look for distant mets
- Also look for suspicious local nodes that can be sampled in EUS
EUS
- Most useful in patients with M0 disease after CT and PET - unnecessary in patients with M1 disease
- Can provide information on T-status and evaluation of regional lymph nodes, although not as accurate for superficial tumours (T1a-T2), where EMR gives better information
- Best use is probably to look for, and then target with FNA, suspicious nodes that may be outside of field and thus change management
  - Nodes are suspicious if they have distinct borders, rounded appearance, hypoechogenic architecture, and size >1cm
EMR can stage suspicious lesions in background of widespread mucosal changes
- Good for T1a-T2 lesions
Laparoscopy
- Stage tumours around GOJ
Bronchoscopy
- Useful in proximal and middle third oesophageal cancers to assess for direct tracheal invasion
For SCC, flexible laryngoscopy and exclusion of synchronous head/neck cancer is recommended

Staging: depth of invasion, lymph node involvement, distant disease, histologic characteristics

Separate staging for SCC and adenocarcinoma
Stage GOJ adenocarcinomas as oesophageal cancer, except for Siewert III tumours (those 2-5cm below GOJ) which are classified as gastric cancers
Oesophageal adenocarcinoma - nodal disease from coeliac axis up to paratracheal region is considered regional disease, while nodal disease outside of those boundaries is considered distant disease
Oesophageal SCC - peri-oesophageal cervical lymphadenopathy is still considered regional disease

Siewert I: tends to move to mediastinal lymph nodes - trans-thoracic en-bloc oesophagectomy with two-field lymphadenectomy
Siewert II: tends to move to coeliac nodes - Ivor Lewis
Siewert III: tends to move to coeliac ndoes - total gastrectomy with trans-hiatal resection of distal oesophagus and extended upper abdominal lymphadenectomy

MANAGEMENT

Initial considerations

Curative or palliative intent
Fitness for surgery
Resectable
Choice of neoadjuvant treatment

General stage-based approach

As of 2024, no consensus on whether to treat SCC and AC any differently
Pre-treatment optimisation is vital
- Nutritional assessment and management, including enteral access if required - NGT through tumour on diagnosis if unable to tolerate pureed/solids. 60% of newly-diagnosed upper GI cancer have malnutrition.
- Management of other comorbidities
Mucosa-only cancer (TisN0M0 and HGD) - intensive surveillance vs endoscopic ablation vs oesophagectomy
- See separate topic under 'Barrett's oesophagus'
- Current best option probably endoscopic in the first instance
T1aN0M0
- Only 2% lymph node mets (not much lymphoid tissue in lamina propria/muscularis mucosae)
- Can usually be treated with EMR
  - Better with lesions <2cm
  - Indicated in nodular or raised Barrett's or other findings suspicious of superficial invasive cancer
  - Resects down to submucosa
  - Does not address the potential for nodal disease (need to assess risk of nodal mets using nomogram above)
- Higher-risk T1a lesions (larger tumours or lymphovascular invasion) or extensive multifocal lesions/ulcerated tumours could be considered for oesophagectomy
T1bN0M0 - oesophagectomy generally best due to abundance of lymphoid tissue in the submucosa
- Up to 20% lymph node metastasis risk - probably should go back and do oesophagectomy if this is the final EMR diagnosis
- However low-risk sm1 may be able to have oesophagus-sparing therapy, with 84% five-year survival rate
  - Needs intense high-dose PPI, H2 blockers, and sucralfat, with 3-monthly endoscopy for the first year
- EMR not considered adequate for sm2 and above due to risk of missing nodal mets
- Generally, give neoadjuvant therapy
T2-4a, any N, M0 ('locally advanced')
- Easy to get locoregional control with oesophagectomy, but rates of death from distant recurrence or metastasis continue to be high
- Neoadjuvant CTX followed by surgery 5-7 weeks later is best strategy
- Restage with CT after CTX
- Clinical T2N0 is somewhat controversial
  - Many have node-positive disease in histopathology after oesophagectomy
  - One strategy is to selectively offer neoadjuvant CTX based on the pretest probability of upstaging (long tumours >3cm, presence of lymphovascular invasion, and high-grade tumours indicate likely to upstage)
T4b or any M1 - palliative - consider palliative chemoradiotherapy
- Endoscopic palliation - self-expanding stents

Neoadjuvant/adjuvant therapy

Consider everything > stage I (non-superficial) cancers for multimodality therapy
- T2-T4
- Nodal involvement with no mets
Regime choices/trials
- MAGIC trial (2006) demonstrated advantage to neoadjuvant chemotherapy
  - NEOAGIS didn't show much difference between MAGIC CTX and CROSS CRT
- CROSS - chemoradiotherapy - weekly carboplatin and paclitaxel for five weeks, with concurrent 41.4Gy radiotherapy in 23 fractions - better tolerated than FLOT, used for patients with comorbidities or poor performance status
  - Landmark trial was the CROSS trial - 366 patients, R0 in neoadjuvant CRTx was 92% vs 69% for up-front surgery, and double the median survival (49 vs 24 months)
- FLOT-4 trial - chemotherapy 5-fluorouacil, folinic acid, oxaliplatin, and docetaxel for three cycles, then surgery, then another three cycles - perhaps this is the best neoadjuvant CTx for adenocarcinoma - can be hard to tolerate for comorbid/frail patients, with mortality 2%
  - ESOPEC 2024 trial showed an advantage to FLOT over CROSS in terms of long-term survival
- CHECKMATE577 - CROSS vs CROSS followed by adjuvant nivolumab immunotherapy - best results for SCC - give as adjuvant if did not achieve complete response - especially good for SCC.
  - Should we also be using it in complete response, to treat microscopic distant disease?
- TOPGEAR - Australian trial which showed no benefit to neoadjuvant radiotherapy
It's not really known what to do with patients who initially have surgery but are then found to have pathological indications for chemoradiotherapy
- May reduce local recurrence in patients with T4 or node positive tumours who didn't have CTX/RTX initially
Chemotherapy
- Usually downstages marginally resectable tumours, allowing improved R0 rates
- Decreases the rate of locoregional recurrence
- Current regimes based cisplatin/carboplatin and 5-fluorouacil/taxanes
- Both adenocarcinoma and SCC patients experience a benefit
- Much better survival with neoadjuvant compared to adjuvant CTX
- Synergistic effect with RTX
Immunotherapy
- Usually nivolumab which is a checkpoint inhibitor with anti-PD-L1 activity
- Start adjuvant immunotherapy 12/52 post op, if indicated based on presence of residual disease
- Duration 1 year
- SCC or adenocarcinoma
- Improved median disease-free survival from 11 to 22 months
Radiotherapy
- 50.4 Gy of radiation used concomitantly with CTX is both a neoadjuvant and potentially definitive dose
Definitive chemoradiotherapy can be curative
- Can do RTx + FOLFOX, CROSS or fluorouracil/cisplatin
- Need surveillance, and salvage therapy if recurrence is found
- 98% of local recurrences occur in first 36 months - suggest vigilant surveillance for this period
- Combination of clinical examination, gastroscopy and CT +/- PET
- SCC is more likely than AC to achieve complete response based on definitive CRT, but there isn't yet consensus on indications for CRT vs surgery in those patients. If complete endoscopic and radiological response has been achieved for oesophageal SCC after CRT, likely appropriate to observe.

Early cancers

Use this diagram to subclassify depth of invasion of superficial cancers
M1, M2 and M3 are reasonable candidates for endoscopic resection, if no evidence of lymphovascular invasion
- If M3 with lymphovascular invasion, treat as T1b
- If recurrence/positive margins, needs oesophagectomy
All submucosal tumours (T1b) should have oesophagectomy with therapeutic lymphadenopathy

Recurrent cancers

Not much evidence for or against salvage oesophagectomy after CRT, but can be done in the knowledge that risks are higher and systemic cure not guaranteed
Resection of oligometastatic disease should be carefully considered, taking disease tempo and ECOG status into account

Palliation

Poor performance status or metastases or unresectable cancers
Checkpoint inhibitors have given good results as palliative CTX
Endoscopic options based on symptoms

Prognosis:

5-year survival:
- Tis 90%
- pT1 75%
- pT2 45%
- pT3 30%
- pT4 15%
Overall 5 year survival 23.7%
Adenocarcinoma
SCC